# Panel: Government Applications of Decision Analysis

Session: SA01
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Thomas A. Edmunds
Chair Address: Lawrence Livermore National Laboratory, Decision Sci. Group, 7000 East Ave. Livermore, Livermore, CA
Chair E-mail: edmunds2@llnl.gov
Chair:
Chair E-mail:

SA01.1 Panel: Government Applications of Decision Analysis
• John Ziagos; Lawrence Livermore National Laboratory, Decision Sci. Group, 7000 East Ave. Livermore, Livermore, CA;
• Gregory S. Parnell; US Military Academy, Dept. of Systems Eng., West Point, NY 10996-1779; fg7526@exmail.usma.edu
• Detlof Von Winterfeldt; University of Southern California, Sch. of Policy/Planning/Dev., Los Angeles, CA 90089; detlof@aol.com
• Thomas A. Edmunds; Lawrence Livermore National Laboratory, Decision Sci. Group, 7000 East Ave. Livermore, Livermore, CA; edmunds2@llnl.gov

Because decisions in the government sector often require consideration of a wide range of non-economic objectives and must be acceptable to a diverse set of stakeholders, they can be more complex than decisions in the private sector. We will describe some of our experiences in introducing decision analysis techniques into government decision processes...

# Representation & Solution of Asymmetric Decision Problems

Session: SA02
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Prakash P. Shenoy
Chair Address: University of Kansas, School of Bus., Summerfield Hall, Lawrence, KS 66045-2003
Chair E-mail: pshenoy@ukans.edu
Chair:
Chair E-mail:

SA02.1 Representing & Solving Asymmetric Decision Problems
• Thomas D. Nielsen; Aalborg University, Dept. of Computer Sci., Frederik Bajers Vej 7C, Aalborg, DK-9220 , Denmark; raistlin@cs.auc.dk
• Finn V. Jensen; Aalborg University, Dept. of Computer Sci., Frederik Bajers Vej 7C, Aalborg, DK-9220 , Denmark; vj@iesd.auc.dk

We present a formal framework called 'asymmetric influence diagrams' based on influence diagrams that allows an efficient representation of asymmetric decision problems. We also give an algorithm for solving asymmetric influence diagrams by decomposing the problem into symmetric subproblems.

SA02.2 Asymmetry in Decision Making in Practice

We address the prevalence, significance and variety of asymmetry in decision-making in practice in a number of applications. In particular, we provide examples from decision problems in project management and credit and risk management and identify asymmetries peculiar to these applications.

SA02.3 Representing Asymmetric Bayesian Decision Problems using Belief Functions

By viewing asymmetry as an uneven numerical specification of probabilities and utilities, a belief functions provide a most natural and compact representation of asymmetric decision problems. It avoids dummy events and acts, and degenerate probabilities and utilities. It also takes full advantage of numerical coalescence.

SA02.4 A Note on Asymmetry in Decision Problems
• Riza Demirer; University of Kansas, Sch. of Bus., Summerfield Hall, Lawrence, KS 66045-2003; riza@ukans.edu
• Prakash P. Shenoy; University of Kansas, School of Bus., Summerfield Hall, Lawrence, KS 66045-2003; pshenoy@ukans.edu

Several methods have been proposed for representing and solving asymmetric decision problems. However, it is not clear whether these methods are capable of representing every asymmetric problem. We study several asymmetric decision problems and examine their representation and solution using existing methods.

# Supply Chain Management

Session: SA03
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Gerard Cachon
Chair Address: University of Pennsylvania, The Wharton Sch., 1300 SH-DH, Philadelphia, PA 19104
Chair E-mail: gpc@mail.duke.edu
Chair:
Chair E-mail:

SA03.1 Beyond Revenue Management: Dynamic Auctioning of Capacity in Overloaded Production Systems
• Rene Caldentey; MIT, Sloan Sch. of Mgmt., E53-343, Cambridge, MA 02139; rcaldent@mit.edu
• Lawrence M. Wein; MIT, Sloan Sch. of Mgmt., Cambridge, MA 02142; lwein@mit.edu

We consider a M/M/1 queue that has a higher arrival rate than service rate. In this setting, auctioning capacity slots to the highest bidders has the potential to significantly increase system revenues, by revealing customers' willingness-to-pay. We analyze this system using queueing theory and auction theory.

SA03.2 Quantifying the Benefits of Supply Chain Visibility
• Bin Yu; Columbia Business School, Grad. Sch. of Bus., 3022 Broadway, Uris Hall 407, New York, NY 10027;
• Fangruo Chen; Columbia Business School, Grad. Sch. of Bus., 3022 Broadway, Uris Hall 407, New York, NY 10027; fc26@columbia.edu

We present a supply chain inventory model with one supplier and one retailer. In one scenario, the supplier makes her production schedule visible to the retailer so the latter can accurately determine the leadtime for each of his orders. In the other, there is no sharing of information. By comparing these scenarios, we begin to understand the value of supply chain visibility.

• Krishnan S. Anand; Northwestern University, Kellogg Grad. Sch. of Mgmt., Evanston, IL 60208; k-anand@nwu.edu

I model the supply chains of 2 firms competing in the same output market. While most previous work has focused primarily on either market-structure (competition) or the firms' internal activities (supply chains), this model studies the interaction between the 2. The results have important strategic implications for example, more information may make a supply chain worse off, even in the absence of incentive problems.

SA03.4 The Vice & Virtue of Channel Stuffing
• Gerard Cachon; University of Pennsylvania, The Wharton Sch., 1300 SH-DH, Philadelphia, PA 19104; gpc@mail.duke.edu

Channel stuffing occurs when a manufacturer encourages a retailer to purchase more stock than needed to cover short run sales, i.e., when the retailer forward buys. This practice has generally been admonished, but this paper finds that channel stuffing may be quite rational.

# Planning & Assignment of Workers

Session: SA04
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: David A. Nembhard
Chair Address: University of Wisconsin, Dept. of IE, 1513 University Ave., Madison, WI 53706-1572
Chair E-mail: nembhard@engr.wisc.edu
Chair:
Chair E-mail:

SA04.1 Worker Training & Assignment in Manufacturing Cells
• Bryan A. Norman; University of Pittsburgh, 1033 Benedum Hall, Dept. of IE, Pittsburgh, PA 15261; banorman@engrng.pitt.edu
• Wipawee Tharmmaphornphilas; University of Pittsburgh, 1048 Benedum Hall, Pittsburgh, PA 15261; witst5+@pitt.edu
• Kim L. Needy; University of Pittsburgh, 1048 Benedum Hall, Pittsburgh, PA 15261; kneedy@engrng.pitt.edu
• Bopaya Bidanda; University of Pittsburgh, 1048 Benedum Hall, Pittsburgh, PA 15261; bidanda@engrng.pitt.edu

We propose a methodology for identifying and assigning the necessary worker skills in manufacturing cells. Then, using these skills, we develop worker training programs and job assignments in order to maximize cell performance. We also consider different training policies and their effect on cell performance.

SA04.2 Queueing Analysis of the Robustness of Skill Chaining Strategies for Cross-Trained Workers
• Mark P. Van Oyen; Loyola University, Dept. of ISOM, 25 East Pearson St., Chicago, IL 60611; mvanoye@luc.edu
• Wallace J. Hopp; Northwestern University, IE/MS Dept., 2145 Sheridan Rd., Tech C210, Evanston, IL 60208-3119; hopp@nwu.edu

We discuss the strategic use of 'D-skill chaining', an approach to worker cross-training appropriate for both serial and parallel production systems. With N stations, N workers are each given D skills. Stochastic queueing models and simulation demonstrate the robustness of the 'opportunity' for performance improvement using dynamic worker assignments.

SA04.3 Work Group Effects on Worker Speed
• Kenneth L. Schultz; Indiana University, Kelley Sch. of Bus., 1309 East 10th St., Bloomington, IN 47401; keschult@indiana.edu
• Richard Garrett; Indiana University, Kelley Sch. of Bus., 1309 East 10th St., Bloomington, IN 47401; rgarrett@indiana.edu
• David A. Nembhard; University of Wisconsin, Dept. of IE, 1513 University Ave., Madison, WI 53706-1572; nembhard@engr.wisc.edu

Recent behavioral research suggests that the speed of a worker is correlated with the speed of others in their work group. It suggests that workers slow down or speed up based on the pace of their co-workers. We use empirical data from a radio manufacturing line to measure and provide evidence of this effect.

SA04.4 Worker Assignments in an Environment of Change
• David A. Nembhard; University of Wisconsin, Dept. of IE, 1513 University Ave., Madison, WI 53706-1572; nembhard@engr.wisc.edu

With increases in make to order and smaller batch sizes workers are often subject to being on the learning curve frequently. We examine methodologies that aid in assigning workers to various tasks in order to improve overall productivity in this environment.

# Analysis & Control of Communications Networks

Session: SA05
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Doug Down
Chair Address: Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205
Chair E-mail: down@isye.gatech.edu
Chair:
Chair E-mail:

SA05.1 Dynamics of a Congestion Pricing Model

We study the use of pricing as a mechanism for allocating bandwidth in communication networks. Starting with a 0/1 pricing scheme proposed by Gibbens & Kelly, we use a distributed control framework to study whether providing more information to users can lead to faster convergence to an efficient allocation of bandwidth in the network.

• David M. McDonald; University of Ottawa, Dept. of Math. & Stats., 585 King Edward Ave., Ottawa, Ontario, K1N 6N5 , Canada; dmdsg@omid.mathstat.uottawa.ca

When the heaviest loaded node in a Jackson network overloads, the other nodes remain stable with larger loads. If, however, the network is modified so that an idle server helps the busy ones then the numbers of customers in the other nodes perform a Brownian excursion.

SA05.3 A Conditional Large Deviation Result for Levy Processes

We look at the way a communications node, driven by certain Levy heavy-tailed -type input, overflows, and identify the optimal path, which, as in similar situations, consists of a large jump, instead of building up linearly. We also discuss some implications of this result.

SA05.4 Asymptotics for Polling Models with Limited Service Policies
• Woojin Chang; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205; woojin@isye.gatech.edu
• Doug Down; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205; down@isye.gatech.edu

We develop expressions for the exact asymptotics of exponential polling models under limited service policies. In addition, we explore the implications of these results for system performance (contrasting with usual measures such as expected waiting times) and buffer allocation problems.

Session: SA06
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Donald W. Hearn
Chair Address: University of Florida, Ctr. for Applied Opt., ISE Dept., Gainesville, FL 32611
Chair E-mail: hearn@ise.ufl.edu
Chair:
Chair E-mail:

SA06.1 Toll Pricing for Variable Demand Traffic Assignment Models
• Mehmet B. Yildirim; University of Florida, Ctr. for Applied Opt., ISE Dept., Gainesville, FL 23611; ybayram@cao.ise.ufl.edu
• Donald W. Hearn; University of Florida, Ctr. for Applied Opt., ISE Dept., Gainesville, FL 32611; hearn@ise.ufl.edu

We propose a variable demand traffic assignment model with side constraints on aggregate flows and demand. Special cases of this model include the elastic demand traffic assignment problem with/without side constraints and the combined distribution-assignment model. We characterize the toll set for the model and then extend the toll pricing framework to the variable demand case.

SA06.2 A Congestion Pricing Scheme on the San Francisco Bay Bridge
• Kara Kockelman; University of Texas, Dept. of Civ. Eng., ECJ 6.9, C1761, Austin, TX 78712; kkockelm@mail.utexas.edu
• Katsuhiko Nakamura; University of Texas, Dept. of Civ. Eng., ECJ 6.9, C1761, Austin, TX 78712;

Daganzo's proposed scheme for reducing congestion via pricing and rationing on a highway with a single bottleneck, assuming socio-economic classes with different values of time, is applied to the San Francisco Bay Bridge; Pareto-improving solutions are sought. The focus is on work trips from the east side of the bridge to downtown San Francisco.

SA06.3 Mixed Multinomial Choice Models for Analyzing Congestion Pricing Strategies

A stated preference survey conducted in the San Francisco Bay area is used to examine the travel behavior responses of Bay Area commuters to different congestion pricing schemes on Bay Area bridges. A mixed multinomial choice formulation, that accommodates correlation in repeated-choices from the same individual, is applied in the analysis.

SA06.4 Evaluation of High Occupancy Toll Lanes using Dynamic Traffic Assignment with a Stochastic Mode Choice Model

HOT lanes provide a new travel alternative that has not been adequately modeled with conventional planning tools. A stochastic mode choice model is incorporated in a dynamic traffic assignment-simulation procedure to analyze the performance and evaluate the effectiveness of alternative dedicated lane configurations and pricing strategies.

# Constructive Methods for Dealing with Special Classroom Issues

Session: SA07
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: James J. Cochran
Chair Address: Louisiana Tech. University, Dept. of Comp. IS & Anlysis, Coll. of Admin. & Bus., Ruston, LA 71272
Chair E-mail: cochran@cab.latech.edu
Chair:
Chair E-mail:

SA07.1 Teaching Modeling across the Business Curriculum
• Christopher J. Zappe; Bucknell University, Dept. of Mgmt., Lewisburg, PA 17837; zappe@bucknell.edu

We explore how the teaching of modeling can be enhanced by making connections with concepts and problems from various disciplines within business administration. Specific illustrations that demonstrate the efficacy of modeling in such areas as accounting, finance, human resources, marketing and operations will be provided.

SA07.2 Extending Instruction Beyond the Quantitative Methods Classroom
• Priscilla Chaffe-Stengel; California State University, 5245 North Backer Ave., MS PB07, Fresno, CA 93740-8001; pchaffe@csufresno.edu

Core courses in quantitative methods remain a significant challenge for business undergraduates. We report the effects of extending instruction beyond traditional class time. Analyses include objective measures of course successes for participants and non-participants and subjective measures of satisfaction. Additional time spent and number of sessions attended are included.

SA07.3 Using the Web to Support a Management Science Course Offered Partly via Distance Learning

We describe the use of Web CT to support a 50% distance learning, 50% face-to-face version of the first MBA course in MS. Organization of the materials, class requirements including homework, exams and terms project and the use of the calendar, chat rooms and private messaging are discussed.

SA07.4 Who Wants to be a Millionaire: The Classroom Edition
• James J. Cochran; Louisiana Tech. University, Dept. of Comp. IS & Anlysis, Coll. of Admin. & Bus., Ruston, LA 71272; cochran@cab.latech.edu

When teaching an introductory Statistics course that meets in long uninterrupted class periods, the author wanted to give students a short pedagogically productive break. Who Wants to be a Millionaire: the Classroom Edition was developed to meet this objective. We explain how the game is administered and give a brief demonstration.

# Financial Engineering

Session: SA08
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Kian Esteghamat
Chair Address: Princeton University, Dept. of ORFE, Princeton, NJ 08544
Chair E-mail: kestegha@princeton.edu
Chair:
Chair E-mail:

SA08.1 Irreversible Investment & Growth Option: Strategy & Valuation

We consider the irreversible investment problem of a firm facing stochastic demands in a continuous time framework. The investment decision has to be made in a finite time horizon, and the value of the investment hinges on the growth options it creates. We develop a quasi-analytic approach to the investment problem. We derive an optimal investment strategy...

SA08.2 A Jump Diffusion Model for Option Pricing with Three Properties: Leptokurtic Feature, Volatility Smile & Analytical Tractability

Two empirical puzzles have got much attention recently: the leptokurtic feature that the return distribution of assets may have a higher peak and 2 (asymmetric) heavier tails than those of the normal distribution and an empirical abnormality called 'volatility smile' in option pricing. To explain them, a jump diffusion model is proposed...

SA08.3 no show

SA08.4 A Boundary Crossing Model of Counterparty Performance

Creditors are familiar with use of net worth as a measure of borrower solvency. This notion can be generalized with a performance state-vector as the indicator of credit quality. We demonstrate a simple, flexible market-based approach for treating counterparty performance in terms of crossings of a state-vector through boundaries.

# Airline Revenue Management

Session: SA09
Date/Time: Sunday 08:30-10:00
Sponsor: Revenue Management Section/Aviation Applications Section
Track:
Cluster:
Room:
Chair: P. V. Krishnarao Pinnamaneni
Chair Address: Enron Corp., 1400 Smith St., Houston, TX 77479
Chair E-mail: pkrishn@enron.com
Chair:
Chair E-mail:

SA09.1 Managing Airline Distribution Channels for Improved Revenue
• E. Andrew Boyd; PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006; aboyd@prosrm.com

The myriad of distribution channels for seat inventory coupled with complex business practices and legacy software systems has led to an extremely complicated environment for RM. We discuss solutions to many of the challenges facing RM today, including alliances and e-distribution channels, with a focus on simplification and improved revenue generation.

SA09.2 Airline Revenue Management with Demand Driven Dispatch

By matching aircraft capacity to demand on a departure by departure level, estimates of unconstrained demand can be verified. Other lessons about RM can also be learned and will be discussed.

SA09.3 Airline Seat Inventory Control with Group Demand
• Youyi Feng; Enron Corp., 1400 Smith St., Houston, TX 77002; fengyy@comp.nus.edu.sg
• Baichun Xiao; Long Island University, Sch. of Bus., 720 Northern Blvd., Brookville, NY 11548; bxiao@liu.edu

We consider an airline seat control model with group demand. Although monotonicity of marginal seat revenue is observed in most airline RM models, it does not explain group discount fares. We show that it is uncertainty and risk consideration that leads to discount fares for group demand.

SA09.4 The Impact of O&D vs. Leg Forecasting on the Demand Unconstraining Process
• Ed Kambour; PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006; ekambour@prosrm.com
• Shankar Sivaramakrishnan; PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006; ssivaramakrishnan@prosrm.com
• E. Andrew Boyd; PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006; aboyd@prosrm.com

Theoretical and simulation results are presented highlighting the differences in forecasting and unconstraining at the O&D and leg levels. The results are applicable to industries other than the airline industry, with proper interpretations on 'leg' and 'O&D'.

# Scheduling Manufacturing/Service Systems

Session: SA10
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: Scheduling
Room:
Chair: Chelliah Sriskandarajah
Chair Address: University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO4.7, Richardson, TX 75083-0688
Chair E-mail:
Chair:
Chair E-mail:

SA10.1 Scheduling Internet Advertising: A Space-Sharing Problem
• Subodha Kumar; University of Texas at Dallas, Sch. of Mgmt., Box 830688, JO 4.7, Richardson, TX 75083-0688; subodha@utdallas.edu
• Varghese S. Jacob; University of Texas, Sch. of Mgmt., PO Box 830688, 2601 North Floyd Rd., Richardson, TX 75083-0688;
• Chelliah Sriskandarajah; University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO4.7, Richardson, TX 75083-0688;

Consider the problems of scheduling ads on Web page, where the space allotted for ads is limited. It has to be decided that which ads will be selected to appear on Web page and how they will be placed, in order to maximize the profit. We develop hybrid heuristics by combining GAs with efficient heuristics.

SA10.2 Scheduling in Openshop Robotic Cells for Productivity Gains
• Inna Drobouchevitch; University of Texas at Dallas, Sch. of Mgmt., Box 830688, JO 4.7, Richardson, TX 75083-0688; innas@utdallas.edu
• Suresh P. Sethi; University of Texas at Dallas, Sch. of Mgmt., Box 830688, Richardson, TX 75083-0688;
• Chelliah Sriskandarajah; University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO4.7, Richardson, TX 75083-0688;
• Jeffrey B. Sidney; University of Ottawa, Faculty of Admin., Ottawa, Ontario, K1N 6N5 , Canada;

Robotic cells are used in repetitive production of identical parts. A robotic cell contains 2 or more robot-served machines. The cycle time is the time to produce a part in the cell. We investigate the productivity advantage of openshop robotic cells to flowshop cells by comparing their cycle times.

SA10.3 Scheduling & Lot Streaming in Machine Shops with No-Wait in Process
• Nicholas G. Hall; Ohio State University, 301 Hagerty Hall, 1775 South College Rd., Columbus, OH 43210-1144; halln@cob.ohio-state.edu
• Gilbert Laporte; University of Montreal, Ecole des HEC, 3000 Chemin Cote-Ste-Catherine, Montreal, Quebec, H3C 3J7 , Canada; gilbert@crt.umontreal.ca
• Esaignani Selvarajah; University of Toronto;
• Chelliah Sriskandarajah; University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO4.7, Richardson, TX 75083-0688;

Consider the problems of minimizing the makespan in no-wait flowshops and 2 machine openshops producing multiple products with detached setup times, using lot streaming. We show that these problems are equivalent to classic TSPs with a pseudopolynomial number of cities. We develop and test computationally efficient heuristics.

SA10.4 Production Scheduling with Cost Variations & Demand Forecast Updates
• Houmin Yan; Chinese University of Hong Kong, Dept. of SE/EM, Ho Sin Hang Engineering Bldg., Hong Kong, Shatin NT, , PR China; yan@se.cuhk.edu.hk
• Sitong Tan; Nankai University, Dept. of Comp. & Systems Sci., Tianjing, 300071 , China;

We study a problem of determining a production sequence and production quantities for multiple products with both demand forecast updates and cost variations. For the unconstrained problem, we discuss special cases that can be solved by some simple sequencing rules. For the problem with capacity constraints, an optimal algorithm is presented to solve the special cases.

# Recent Advances in Airline Optimization

Session: SA11
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: Integer Programming
Room:
Chair Address: Transport Dynamics Inc., 103 Carnegie Ctr., Ste. 317, Princeton, NJ 08540
Chair E-mail: jarrah@transdynamics.com
Chair:
Chair E-mail:

SA11.1 Advances in the Airline Crew Pairing Problem
• Amy M. Cohn; MIT, Dept. of Civil & Environ. Eng., Rm. 1-229, Cambridge, MA 02139-4307; amycohn@mit.edu
• Cynthia Barnhart; MIT, OR Ctr., Rm. 1-229, 77 Massachusetts Ave., Cambridge, MA 02139; cbarnhar@mit.edu

The airline crew pairing problem is an integral part of airline planning. Although the basic model is well established in the literature, there is significant room for enhancement to improve solution quality and computational performance and to expand the model scope. We present our models, algorithms and computational results.

SA11.2 Optimizing Personnel Training & Deployment at Airlines
• Benjamin G. Thengvall; CALEB Technologies Corp., 9130 Jollyville Rd., Austin, TX 78759; ben@calebtech.com
• Michael F. Arguello; CALEB Technologies Corp., 9130 Jollyville Rd., Austin, TX 78759; arguello@calebtech.com
• Gang Yu; University of Texas, Dept. of MSIS, Gras. Sch. of Bus. Admin., Austin, TX 78712-1175; yu@uts.cc.utexas.edu

Deploying a work force to meet an airline's business plan requires a coordinated management approach. Intelligent timing of new hires and transitional training for pilots and flight attendants can provide significant revenue enhancing opportunities. We will present an overview of the problem and a framework for its solution.

SA11.3 A New Approach to Aircraft Rotation Planning
• Dimitry Keselman; Idmon Corporation, 2141 Powers Ferry Rd., Ste. 310, Marietta, GA 30067; dkeselman@optimization.com
• Roy E. Marsten; Idmon Corporation, 2141 Powers Ferry Rd., Ste. 310, Marietta, GA 30067;

One of the most important and desirable features of an aircraft rotation is the ability to fly all fleet planes on the same route (one sub-rotation). We discuss some graph and linear programming models employed to ensure the one sub-rotation property during the fleeting and rotation building phases.

SA11.4 Airline Crew Scheduling with Time Windows & Plane Count Constraints

Airline planning consists of several problems that are currently solved separately. We address a partial integration of schedule planning, aircraft routing and crew scheduling. In particular, we provide more flexibility for crew scheduling while maintaining the feasibility of aircraft routing by adding plane count constraints to the crew scheduling problem. We assume that the departure times of flights have not yet been fixed...

# Strategic Issues in Supply Chain Management

Session: SA12
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: Logistics & Supply Chain Management
Room:
Chair: Jovan Grahovac
Chair Address: Tulane University, Freeman Sch. of Bus., 7 McAlister Dr., New Orleans, LA 70118
Chair E-mail: jovan.grahovac@tulane.edu
Chair:
Chair E-mail:

SA12.1 Capacity & Backlog Management in Service-Oriented Supply Chains

Although there is much research on managing inventory in supply chains, very little is available on managing capacity in the absence of inventory, which is a typical problem in service supply chains. We begin to fill this void by analyzing supply chain capacity management strategies enabled by electronic business

SA12.2 Entry & Competition in Multi-Echelon Supply & Distribution Networks
• Scott Carr; UCLA, Anderson Sch. of Mgmt., Los Angeles, CA 90095;
• Uday S. Karmarkar; UCLA, Anderson Sch., 110 Westwood Plaza, Box 951481, Los Angeles, CA 90095-1481; uday.karmarkar@anderson.ucla.edu

We study cost structure and competition in supply networks for multi-echelon assembly and distributive structures. Entry at any tier incurs a fixed cost entrants then choose production quantities. The entry game is characterized as a Nash equilibrium while post entry competition is modeled as a 'successive Cournot oligopoly'.

SA12.3 withdrawn - author request of 9/11
• Kyle D. Cattani; University of North Carolina, Kenan-Flagler Bus. Sch., CB 349, Chapel Hill, NC 27599-3490; kyle_cattani@unc.edu
• Ely Dahan; MIT, Sloan Sch., 38 Memorial Dr., E56-323, Cambridge, MA 02142; edahan@mit.edu
• Glen Schmidt; Georgetown University, McDonough Sch. of Bus., G-4 Old North, 37th & O St. NW, Washington, DC 20057; schmidtg@gunet.georgetown.edu

SA12.4 An Inventory Model with a Two-Tiered Service Level Model for a 'Bricks & Clicks' Retailer
• Doug Thomas; Pennsylvania State University, Smeal Coll. of Bus. Admin., University Park, PA 16802-3005;

While the growth in e-tailing and consumer-direct delivery certainly puts traditional retailers at risk, these retailers still offer value to consumers through better service, including easier returns/exchanges and permitting the customer to physically examine the product. Companies choosing to augment their physical presence with an on-line presence, a 'bricks and clicks'' strategy, attempt to capture the best of both worlds...

# Information Coordination in Logistics

Session: SA13
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: Manufacturing & Logistics
Room:
Chair: Suvrajeet Sen
Chair Address: University of Arizona, Dept. of SIE, Tucson, AZ 85721
Chair E-mail: sen@sie.arizona.edu
Chair:
Chair E-mail:

SA13.1 Defense Distribution Center DC Supply Chain Network: Factors that Impede/Influence the Flow of Information
• Rekha S. Pillai; Oak Ridge National Laboratory, Oak Ridge, TN 37831;
• Glen Harrison; Oak Ridge National Laboratory, Oak Ridge, TN 37831;
• Cheng Liu; Oak Ridge National Laboratory, Oak Ridge, TN 37831;
• Michael R. Hilliard; Oak Ridge National Laboratory, Bldg. 4500N, MS 6270, Oak Ridge, TN 37831-6270;
• Mark Entner; Defense Logistics Agency, OR & Resource Analysis Office, Richmond, VA 23297-5082;

We analyze the DDC supply chain network. A main objective is to improve asset visibility across DDC supply chain to improve customer confidence, meet customer requirements, and reduce inventory and supply chain cost. We identify factors that impede visibility and determine requirements for DDC Supply Chain Network to improve information flow and increase visibility.

SA13.2 Product Tracking for Improved Service in a Global Supply Chain Network
• Pius J. Egbelu; Iowa State University, Dept. of IMSE, 2019 Black Engineering, Ames, IA 50011-2164; pegbelu@iastate.edu

Many assembled products in today's market are built from components and subassemblies made by different manufacturers. These manufacturers often don't have the same manufacturing capabilities nor do they share similar customer service policies. Furthermore, the quality of the end product reflects the strengths and weaknesses of all the manufacturers whose components make up the product. Product coding for ease of tractability and improved customer service will be discussed.

SA13.3 Time-Scale Coordination in Integrated Organizations

Organizations consisting of multiple units often make decisions in different time scales: aggregate production planning decisions are made on a much longer time scale, than dispatching, routing and scheduling decisions. Closing the loop between these decisions plays an important role in coordinating different units within an organization. We will outline a framework within which such coordination can be modeled and a decentralized control approach developed.

# EPD Mini-Cluster: Web/IT Driven Product Development I

Session: SA14
Date/Time: Sunday 08:30-10:00
Track:
Cluster: Product Development
Room:
Chair: Vish V. Krishnan
Chair Address: University of Texas, Dept. of Mgmt., CBA 4.202, Austin, TX 78712
Chair E-mail: krishnan@mail.utexas.edu
Chair:
Chair E-mail:

SA14.1 Design Automation Usage Patterns during Complex Electro-Mechanical Developpment
• Nitin Joglekar; Boston University, Sch. of Mgmt., 595 Commonwealth Ave., Boston, MA 02215; joglekar@bu.edu
• Daniel E. Whitney; MIT, Tech., Policy & Ind. Develop., Cambridge, MA 02139; dwhitney@mit.edu

Our study of automation usage patterns in the aerospace industry shows that rising design complexity and linkages with supply chain management issues influence the structure of the development process. We document how the development time is expended and show that the IT/infrastructure development consumes a significant amount of resources.

SA14.2 R&D Performance Measures that are Linked to Strategy
• Christoph H. Loch; INSEAD, Blvd. de Constance, Fountainbleau Cedex, 77305 , France; christoph.loch@insead.fr
• Stefan Tapper; University of Witwatersrand, Johannesburg, , South Africa;

Companies often struggle to assess R&D performance. No widely accepted performance measurement system for R&D exists. We show how the Research Group of the diamond producer GemStone developed a research performance measurement system that is appropriate for the risky projects typical in R&D and which supports business strategy. The process derives operative measures for R&D from the company strategy...

SA14.3 Role of Information Technology in Product Development: Implications for Problem-Solving & Knowledge Creation
• Stefan Thomke; Harvard Business School, Morgan Hall T63, Soldiers Field, Boston, MA 02163;

With pressure to conduct product development faster and more efficiently, firms are increasingly turning to new technologies such as computer simulation. I will propose that these new technologies can fundamentally change problem-solving and knowledge creation that will require firms to rethink how they manage and organize their development activities.

# Inventory & Coordination

Session: SA15
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: Quantitative Models in Supply Chain Management
Room:
Chair: Lingxiu Dong
Chair Address: Washington University, Olin Sch. of Bus., 1 Brookings Dr., CB 1133, St. Louis, MO 63130-4899
Chair E-mail: dong@olin.wustl.edu
Chair:
Chair E-mail:

SA15.1 Implementing an Inventory Management Strategy with Retailer Cooperation

We present the development, analysis and use of mathematical models in an actual implementation when a supplier of RTA furniture has access to her retailer's inventory-forecasting-promotion-ordering strategy.

SA15.2 Vertical & Horizontal Pricing under Uncertainty
• Nils Rudi; University of Rochester, Simon Sch. of Bus., Rochester, NY 14627; rudi@simon.rochester.edu
• David F. Pyke; Dartmouth College, The Tuck Sch., Hanover, NH 03755-1798;

We consider a scenario where a manufacturer sells to independently owned retailers. These retailers use lateral transshipments for risk pooling purposes. We examine how the use of transshipments affects the manufacturer and the system performance.

SA15.3 Implications of a Pooled Return Policy

Single-product returns policies allow return of a fraction of the order for a product. We consider a multi-product returns (pooled) policy in which the total returns are limited to a fraction of the total order across all products, comparing the profit and decisions to that of the single-product policy.

SA15.4 Selling a Build-to-Order Manufacturer: A Supplier's Pricing Problem
• Lingxiu Dong; Washington University, Olin Sch. of Bus., 1 Brookings Dr., CB 1133, St. Louis, MO 63130-4899; dong@olin.wustl.edu
• Nils Rudi; University of Rochester, Simon Sch. of Bus., Rochester, NY 14627; rudi@simon.rochester.edu

Introducing component commonality across brands affects the profit sharing in a system where a subcontractor purchases components from multiple suppliers and assembles to order for multiple brands. The selling behavior (component wholesale price) of the component suppliers when they are independent or integrated reveals opportunities in collaboration across brands.

# Multiple Objective Programming

Session: SA16
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: MCDM
Room:
Chair: Minghe Sun
Chair Address: University of Texas, Coll. of Bus., Divison of Mktg. & Mgmt., San Antonio, TX 78249-0634
Chair E-mail: msun@utsa.edu,, msun@lonestar.utsa.edu
Chair:
Chair E-mail:

SA16.1 withdrawn - chair request of 9/7
• James P. Ignizio; University of Virginia, Sch. of Eng. & Applied Sci., PO BOx 400747, Charlottesville, VA 22904-4747;

SA16.2 Nonlinear Goal Programming using Genetic Algorithms
• Mitsuo Gen; Ashikaga Institute of Technology, Dept. of IE, Ashikage, 326-8558 , Japan; gen@ashitech.ac.jp

The use of GAs to improve results of nonlinear goal programming analysis is presented.

SA16.3 Finding Integer Efficient Solutions for Multiple Objective Network Programming Problems

An algorithm for finding integer efficient solutions for multiple objective network programming problems is proposed and implemented. The algorithm may find all integer efficient solutions or the one closest to the ideal point in the neighborhood of a fractional efficient solution obtained by solving an augmented weighted Tchebycheff network program.

SA16.4 Data Mining in Credit Card Portolio Management: A Multiple Criteria Decision Making Approach
• Yong Shi; University of Nebraska, Coll. of IS&T, Omaha, NE 68182-0392; yshi@unomaha.edu

A multiple criteria linear programming approach has been identified as the alternative technology to predict cardholder future behavior in credit card portfolio management. Testing a small development sample indicated that this approach is fully controlled, and a sample size of 3000 is suitable for robustness. The method works well with multiple groups.

# Theory & Computation

Session: SA17
Date/Time: Sunday 08:30-10:00
Track:
Cluster: Network Flows
Room:
Chair: Jane L. Snowdon
Chair Address: IBM TJ Watson Research Center, PO Box 218, Rte. 134, Yorktown Heights, NY 10598-0218
Chair E-mail: snowdonj@us.ibm.com
Chair:
Chair E-mail:

SA17.1 A New Column Generation Approach for Airline Crew Scheduling
• Tina L. Shaw; Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332; tshaw@isye.gatech.edu
• Ellis L. Johnson; Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205; ellis.johnson@isye.gatech.edu

Shortest path subproblems have been successfully used to dynamically generate columns for airline crew scheduling problems. We present a new shortest path network and hybrid column generation approach for fast solutions to the LP relaxations along with a B&P method that yields good IP solutions.

SA17.2 Column Generation for Airline Scheduling Problems
• Jane L. Snowdon; IBM TJ Watson Research Center, PO Box 218, Rte. 134, Yorktown Heights, NY 10598-0218; snowdonj@us.ibm.com
• Francisco Barahona; IBM Watson Research Center, PO Box 218, Yorktown Heights, NY 10598; baranon@us.ibm.com
• John J. Forrest; IBM TJ Watson Research Center, PO Box 218, Yorktown Heights, NY 10598; jjforre@us.ibm.com
• Laszlo Ladanyi; IBM Research, TJ Watson Research Ctr., PO Box 218, Rte. 134, Yorktown Heights, NY 10598; ladanyi@us.ibm.com

Airline scheduling at major airlines involves optimizing over many trillions of variables. Managing the problem size requires effective column generation. We explore 2 schemes for column generation, one based on subgradient duals and the other motivated by a quadratic programming formulation. We conclude with

SA17.3 A Second-Stage Network Recourse Problem in Stochastic Airline Crew Scheduling
• Joyce Yen; University of Michigan, Dept. of IOE, 1205 Beal Ave., Ann Arbor, MI 48109-2117; jyen@engin.umich.edu
• John R. Birge; Northwestern University, McCormick Sch. of Engineering, 2145 Sheridan Rd., Evanston, IL 60208-3100; jrbirge@nwu.edu

We model the stochastic airline crew scheduling problem as a 2-stage stochastic program in which the first stage is a set covering problem and the second stage is a network recourse problem. We discuss results from various algorithms to solve this problem and present results on generating disruption scenarios.

# Electronic Commerce

Session: SA18
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: E-Commerce
Room:
Chair: Gary J. Koehler
Chair Address: University of Florida, Dept. of DIS, Box 117169, Coll. of Bus., 351 Stuzin Hall, Gainesville, FL 32611-7169
Chair E-mail: koehler@ufl.edu
Chair:
Chair E-mail:

SA18.1 EDI, XML & the Myth of Semantic Transparency

We review foundational principles of (semantic) communication, and in light of them, the first-trade problem in EDI. We discuss how XML can address the problem and argue that its capacities to do so are severely limited. We conclude with discussion of how to solve the problem, giving illustrative examples.

SA18.2 Does Location Matter in an E-Commerce World? Application of Location Theory & Spatial Economics to Consumer-Based E-Commerce
• Brian E. Mennecke; Iowa State University, Dept. of MIS, 300 Carver Hall, Ames, IA 50011;
• Troy J. Strader; Iowa State University, Dept. of MIS, 300 Carver Hall, Ames, IA 50011-2063; tstrader@iastate.edu

We apply location theory, which is a branch of spatial economic theory, to marketing behavior in a business-to-consumer e-commerce environment with the goal of developing a framework for understanding the role of location in both e-commerce and brick-and-mortar transactions.

SA18.3 no show
• Ram Gopal; University of Connecticut, Sch. of Bus., U-41, 368 Fairfield Rd., Storrs, CT 06269;
• Zhiping Walter; University of Connecticut, Sch. of Bus., U-41 IM, 368 Fairfield Rd., Storrs, CT 06269; zwalter@sba.uconn.edu
• Arvind Tripathi; University of Connecticut, U-41 OP, 368 Fairfield Rd., Storrs, CT 06269; tripathi@sba.uconn.edu

SA18.4 no show
• Erik Rolland; University of California, AGSM, Riverside, CA 92521;
• John H. Gerdes, Jr.; University of California, AGSM, Riverside, CA 92521; john.gerdes@ucr.edu

# Strategic Management of Technology

Session: SA19
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: Technology Innovations & Operations
Room:
Chair: Alan Pearson
Chair Address: Manchester Business School, R&D Res. Ctr., Booth St. West, Manchester, M15 6PB , UK
Chair E-mail: a.pearson@fs2.mbs.ac.uk
Chair:
Chair E-mail:

SA19.1 Strategic Management of New Technology

There has been much debate about the timescales required to exploit scientific and technological developments. Are they shortening or not? The answer to this question must be contingent upon the type of technology and the market environment. We discuss a recent research project in which we made an in depth study of virtual reality...

SA19.2 Requirements for Capability Maturity Models to Improve Communication in Virtual Organizations

Current capability maturity models have been developed to improve performance of development projects. Capabilities at different maturity levels have been defined for projects to perform in a controlled way, for organizations to facilitate their projects and for the people to improve their way of working. Capability maturity models are still mainly internally focused...

SA19.3 Integrated Design Capabilities & Technology Management

In order to keep up with the new product development race that occurs in many industries today, the innovative organization has to become much more adaptive in responding to technological and market changes. In order to achieve this, innovation management has to develop an integrated design capability that allows for fast product experimentation and adaptation...

SA19.4 Information & Communication Technology Divergence in the Context of Information & Communication Technology Convergence

ICT has been sufficiently developed to become another commodity infrastructure but there is more to it. It is argued that the technical dimension will not really be a key factor in future development and that further growth, development and change is coming from the perceived potential in novel and innovative human-directed applications of the technical infrastructure. This shift from ICT to ITC is explored.

# Extensions & Applications of Heuristic Search

Session: SA20
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: Tabu & Scatter Search
Room:
Chair: David L. Woodruff
Chair Address: University of California, GSM, Davis, CA 95616
Chair E-mail: dlwoodruff@ucdavis.edu
Chair:
Chair E-mail:

SA20.1 Metaheuristics & their Applications in Transportation & Logistics Problems
• Buyang Cao; ESRI, Inc., 380 New York St., Redlands, CA 92373;

Metaheuristics have recently been widely applied to various difficult optimization problems due to their efficiency. We specifically use tabu search techniques together with other search strategies, i.e., restart, cost oscillation, etc., to solve optimization problems from transportation and logistics area. We demonstrate the economic benefit achieved by methaheuristics.

SA20.2 A Blackboard Architecture Applied to Maximum Likelihood Clustering
• David L. Woodruff; University of California, GSM, Davis, CA 95616; dlwoodruff@ucdavis.edu
• Torsten Reiners; Technical University of Braunschweig, Braunschweig, , Germany;

We describe an algorithm that provides significant improvements in the ability to perform maximum likelihood clustering. It combines heuristic search strategies via a blackboard architecture to obtain good starting points for an estimator defined as a continuous optimization problem. The architecture is novel because it exploits connections between continuous and discrete spaces to facilitate storage, communication and exploitation of heuristic search information...

SA20.3 Using Explicit Memory in Restart Methods
• Arne Lokketangen; Molde College, Britvegen 2, Molde, 6400 , Norway;
• Fred W. Glover; University of Mississippi, Hearin Ctr. for Enterprise Sci, Sch. of Bus. Admin., University, MS 38677; fglover@bus.olemiss.edu

Increasing attention has recently been given to restart methods, both for constructive and local search heuristics. We look at ways to reduce the customary reliance on randomization in such methods, relying instead on explicit adaptive memory processes. We illustrate the mechanisms on a portfolio of satisfiability.

# Session I

Session: SA21
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: SICUP
Room:
Chair: Gerhard Waescher
Chair Address: Martin-Luther-Universitaet Halle, Wirtschaftswissenschaften, Produktion & Logistik, Halle, Saale, D-06099 , Germany
Chair E-mail: waescher@wiwi.uni-halle.de
Chair:
Chair E-mail:

SA21.1 Uses of the 1-Dimensional Gilmore-Gomory LP Cutting Stock Algorithm

We explore the ways in which the Gilmore-Gomory LP algorithm for solving one-dimensional cutting stock problems can be used to help solve a variety of cutting and packing problems. The examples range from simple bin packing to 3-dimensional trailer loading.

SA21.2 1-Dimensional Cutting: A State-of-the-Art Survey
• Gerhard Waescher; Martin-Luther-Universitaet Halle, Wirtschaftswissenschaften, Produktion & Logistik, Halle, Saale, D-06099 , Germany; waescher@wiwi.uni-halle.de

I will review advances which have been achieved during the last decade in one-dimensional cutting. Not only problems of the cutting stock bin-packing type will be considered, but also problems from other classes which are usually not covered in the area of cutting and packing.

# Resource Utilization in Health Care

Session: SA22
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Michael W. Carter
Chair Address: University of Toronto, Dept. of MIE, 5 King's College Rd., Toronto, Ontario, M5S 3G8 , Canada
Chair E-mail: carter@mie.utoronto.ca
Chair:
Chair E-mail:

SA22.1 Assigning Anaesthetists to Operating Rooms via Multi-Dimensional Analysis
• John T. Blake; DalTech., Dept. of IE, PO Box 1000, Halifax, Nova Scotia, B3J 2X4 , Canada; john.blake@dal.ca
• Sandi Eaves; Dalhousie University, Dept. of IE, PO Box 1000, Halifax, Nova Scotia, B3J 2X4 , Canada;

The Queen Elizabeth Health Sciences Centre, located in Halifax, is Atlantic Canada's largest hospital. We discuss a model for automating the current process of scheduling anaesthetists manually to more than 40 operating rooms each day. The model is a standard assignment problem in which the 'cost' of assigning an anaesthetist to a room is determined from a multi-dimensional penalty function.

SA22.2 withdrawn - chair request of 10/4
• Sophie D. Lapierre; Ecole Polytechnique, CRT, CP 6179, Succ. Centre-ville, Montreal, Quebec, H3C 3A7 , Canada; sophiel@crt.umontreal.ca
• Marko Blais; Ecole Polytechnique, CRT, CP 6179, Succ. Centre-ville, Montreal, Quebec, H3C 3A7 , Canada; mblais@crt.umontreal.ca

SA22.3 Managing Home Care Resource Utilization & Waiting Lists
• Michael W. Carter; University of Toronto, Dept. of MIE, 5 King's College Rd., Toronto, Ontario, M5S 3G8 , Canada; carter@mie.utoronto.ca
• Rebecca Chan; University of Toronto, Dept. of MIE, 5 King's College Rd., Toronto, Ontario, M5S 3G8 , Canada; yyrebecca@hotmail.com
• Edmon Chung; University of Toronto, Dept. of MIE, 5 King's College Rd., Toronto, Ontario, M5S 3G8 , Canada; edmonchung@hotmail.com
• Linda M. Lakats; University of Toronto, Dept. of MIE, 5 King's College Rd., Toronto, Ontario, M5S 3G8 , Canada; lakats@interlog.com

In Ontario, home care services are delivered by regional centres, Community Care Access Centres. All requests for home care are funneled through a single agency. With acute care, hospital bed closures and an emphasis on shorter length of stay, there has been a dramatic increase in the pressure on home care delivery over the past 5-10 years...

SA22.4 withdrawn - chair request of 9/22
• Kevin J. Leonard; University of Toronto, Dept. of Health Admin., 2 Queen's Park Crescent West, Toronto, Ontario, M5S 1A8 , Canada; k.leonard@utoronto.ca

# CANCELLED: Web-Based Simulation

Session: SA23
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Arnold H. Buss
Chair E-mail: abuss@nps.navy.mil
Chair:
Chair E-mail:

SA23.2 withdrawn - chair request of 10/30
• Ronald F. Woodaman; Naval Postgraduate School, OR Dept., Monterey, CA 93943-5000; rfwoodam@nps.navy.mil

SA23.3 withdrawn - chair request of 10/30
• Arnold H. Buss; Naval Postgraduate School, OR Dept., Monterey, CA 93943-5000; abuss@nps.navy.mil

# Facilities Design

Session: SA24
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: Material Handling
Room:
Chair: Rajan Batta
Chair Address: SUNY, Dept. of IE, 342 Bell Hall, North Campus, Buffalo, NY 14260-2050
Chair E-mail: batta@acsu.buffalo.edu
Chair:
Chair E-mail:

SA24.1 Placement of New Facilities under a Mixed Planar/Network Representation of Facility Layout

We provide a mixed planar/network representation for facility layout where machines, cells or departments are planar entities of finite size and the material flow between these entities occur on material handling paths constituting a network structure. We then consider 2 situations for the placement of new facilities where perturbations to the existing network are either allowed or not.

SA24.2 Design of Zone-Based Tandem Layout for Automated Guided Vehicle Systems
• Wooyeon Yu; Iowa State University, Dept. of IE/MS, Ames, IA 50010;
• Pius J. Egbelu; Iowa State University, Dept. of IMSE, 2019 Black Engineering, Ames, IA 50011-2164; pegbelu@iastate.edu

Apartitioning algorithm for a tandem AGV system based on the concept of zones as against loops is developed. The algorithm simultaneously determines transfer points. A performance comparison between the conventional and the zone-based tandem AGVS was undertaken. With respect to system performance under the same operating conditions is presented.

SA24.3 Improved Integrated Scheduling & Logistics in an Assembly Environment using Concepts from Column Generation

A new heuristic based on ideas from column-generation is developed to solve the integrated AGV-workcenter scheduling problem, which is NP-hard. The master problem selects the sequence of operations and their assignment to multiple replicates so that the makespan is minimized. The subproblem generates new sequences of operations with an attempt to minimize reduced cost...

# Nonlinear Programming Algorithms for Semidefinite Programming

Session: SA25
Date/Time: Sunday 08:30-10:00
Track:
Cluster: Nonlinear Programming
Room:
Chair: Renato D. C. Monteiro
Chair Address: Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332
Chair E-mail: monteiro@isye.gatech.edu
Chair:
Chair E-mail:

SA25.1 Solving Semidefinite Programs via Nonlinear Programming: Theory & Algorithms

We show that a large class of linear and/or nonlinear SDP problems can be converted into nonlinear optimization problems over 'orthants' of the form $\Re^n_{++} \times \Re^N$, where $n$ is the size of the matrices involved in the problem.

SA25.2 A Nonlinear Programming Algorithm for Solving Semidefinite Programs via Low-Rank Factorization

We discuss an algorithm for solving SDPs in which the semidefiniteness of the matrix variable is maintained via an implicit low-rank factorization. The rank of the factorization is chosen minimally so as to enhance computational speed. Strong computational results are reported.

SA25.3 A Nonlinear Programming-Based Interior-Point Method for Semidefinite Programs
• Robert J. Vanderbei; Princeton University, Dept. of OR & Civil Eng., Princeton, NJ 08544; rvdb@princeton.edu
• Hande Y. Benson; Princeton University, Dept. of Civil Eng. & OR, Dept. of OR & Financial Eng., Princeton, NJ 08544; hyurttan@princeton.edu

We show how to formulate SDP problems as smooth convex NLP problems and describe modifications to a generic interior point method for NLP to make it applicable to SDP. Preliminary computational results will be given.

# Modeling Queues with Power-Tailed Interarrival and/or Service Distributions: In Honor of Carl Harris

Session: SA26
Date/Time: Sunday 08:30-10:00
Type: Invited
Track:
Cluster: Stochastic Models & Applications
Room:
Chair: Donald Gross
Chair Address: George Mason University, Dept. of Systems Eng. & OR, Fairfax, VA 22030-4444
Chair E-mail: dgross@gmu.edu
Chair:
Chair E-mail:

SA26.1 Dedication of Session to Carl M. Harris, 1940-2000
• Saul I. Gass; University of Maryland, 8809 Maxwell Dr., Potomac, MD 20854-3123; sgass@rhsmith.umd.edu
• Donald Gross; George Mason University, Dept. of Systems Eng. & OR, Fairfax, VA 22030-4444; dgross@gmu.edu

Carl's role in putting this session together and highlights and reminiscences of his eminent career in OR are presented.

SA26.2 G/M/c Queues with Power-Tailed Interarrival Times & a Level Crossing Computational Method for their Analysis
• Percy H. Brill; University of Windsor, Dept. of MS, Math & Stats., Faculty of Bus., Windsor, Ontario, N9B 3P4 , Canada; brill@uwindsor.ca
• Carl M. Harris; George Mason University, Sch. of IT & Eng., Fairfax, VA 22030;

We analyze the waiting time in G/M/c queues having power-tailed interarrival times, analytically using a level crossing method. This method provides numerical results when interarrival times have Pareto, folded Cauchy or log-normal distributions. Simulations confirm the results. Applications to Internet traffic are discussed.

SA26.3 A Transform Approximation Method & its Use in Analyzing Queues with Heavy-Tail Arrival or Service Distributions
• Martin J. Fischer; Mitretek Systems, Telecomm. & Networking, 7525 Colshire Dr., McLean, VA 22101-7400; mfischer@mitretek.org
• Carl M. Harris; George Mason University, Sch. of IT & Eng., Fairfax, VA 22030;

We present and discuss the use of a TAM in the analysis of queues that have heavy- or long-tailed arrivals or service distributions. Computational experiences with TAM and performance improvement techniques are discussed. Comparisons with simulation, level crossing and other methods are presented.

SA26.4 Simulation of Power-Tail Distributions in Queueing Probems
• Donald Gross; George Mason University, Dept. of Systems Eng. & OR, Fairfax, VA 22030-4444; dgross@gmu.edu
• Carl M. Harris; George Mason University, Sch. of IT & Eng., Fairfax, VA 22030;
• Denise M Masi; Mitretek Systems, 7525 Colshire Dr., McLean, VA 22102-7400; dmasi@mitretek.org

Using power-tail distributions in discrete-event simulation modeling of Internet congestion (often necessary in the absence of analytical results) can be fraught with difficulty, especially when dealing with high coefficients of variation. We illustrate these difficulties and compare simulation to the numerical procedures of the previous 2 papers.

# Panel: The Academic Job Search for PhD Students

Session: SA27
Date/Time: Sunday 08:30-10:00
Type:
Track:
Cluster:
Room:
Chair: Russell D. Meller
Chair Address: Virginia Tech., Dept. of ISE, 250 New Engineering Bldg., Blacksburg, VA 24061
Chair E-mail: rmeller@vt.edu
Chair:
Chair E-mail:

SA27.1 Panel: The Academic Job Search for PhD Students

Former academic search chairs will provide PhD students with an outline of the academic job search, i.e., application, interview, offer process, etc. They will also provide their guidance on some of the dos and don'ts.

# The Internet & Social Science Applications

Session: SA28
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Myron Hatcher
Chair Address: California State University, Dept. of IS/DS, Craig Sch. of Bus., Fresno, CA 93740
Chair E-mail: myron_hatcher@csufresno.edu
Chair:
Chair E-mail:

SA28.1 Ethics, Cultures & Generations: Is there a Difference?
• Gerald Jones; California State University, Dept. of IS/DS, Craig Sch. of Bus., Fresno, CA 93740; geraldj@csufresno.edu

We report on the results of research conducted to determine if the generational and cultural differences noted by previous researchers affects decision making. Data were collected from academics on 2 continents representing different cultures. Results provide insight into how important culture and generational affects are in making ethics impacted decisions.

SA28.2 Using Bootstrap Financing Techniques to Fund New & Small Business Firms
• Amir A. Jassim; California State University, Dept. of Finance & Bus. Law, Craig Sch. of Bus., Fresno, CA 93740; amirj@csufresno.edu

Many new entrepreneurs and small business owners find the prospect of raising capital to be a daunting task and a major stumbling block in starting or expanding their existing business. They may have excellent ideas about a new product/service or improving an existing one but have little knowledge about money and finance. We explain the various bootstrap financing techniques...

SA28.3 Internet Potential Impact on Health Care

The potential impact of Internet and Intranets will be discussed. Results from a national survey of acute care hospitals will indicate potential directions.

SA28.4 Design Implications of the Internet on Medical Software
• Thaddeus W. Usowicz; California State University, School of Bus., San Francisco, CA 94127-2716; usowicz@sfsu.edu

The Internet has serious implications for different user interfaces for health care. Realization and ease-of-use has brought about more knowledgeable patents and misdirected parents. At the same time, health care personals have successfully integrated information in practice.

# Auctions & Bidding in the Electric Power Industry

Session: SA29
Date/Time: Sunday 08:30-10:00
Track:
Cluster: Bidding
Room:
Chair: Shmuel S. Oren
Chair Address: University of California, Dept. of IE&OR, 4135 Etcheverry Hall, Berkeley, CA 94720
Chair E-mail: oren@ieor.berkeley.edu
Chair:
Chair E-mail:

SA29.1 Strategic Bidding under Uncertainty in a Competitive Electricity Market
• Alvaro Baillo; Universidad Pontificia Comillas, Inst. de Investigacion Tec., Albert Aguilera 23, Madrid, 28015 , Spain; alvaro.baillo@iit.upco.es
• Mariano Ventosa; Universidad Pontificia Comillas, Inst. de Investigacion Tec., Alberto Aguilera 23, Madrid, 28015 , Spain; mariano.ventosa@iit.upco.es
• Michel Rivier; Universidad Pontificia Comillas, Inst. de Investigacion Tec., Alberto Aguilera 23, Madrid, 28015 , Spain;
• Andres Ramos; Universidad Pontificia Comillas, Inst. de Investigacion Tec., Alberto Aguilera 23, Madrid, 28015 , Spain;

Bidding into the electricity wholesale organized markets turns out to be a very complex matter for generation companies. We have developed a systematic and automatic procedure to deal with the sort-term day-ahead 24-hour bidding decision making. It consists of a probabilistic optimization tool oriented to the construction of profit-maximizing hourly offer curves.

SA29.2 Design of Auctions for Hierarchically Substitutable Ancillary Services
• Shmuel S. Oren; University of California, Dept. of IE&OR, 4135 Etcheverry Hall, Berkeley, CA 94720; oren@ieor.berkeley.edu

The California ISO procures 4 types of reserves to ensure system reliability: regulation, spinning reserves, nonspinning reserves and replacement reserves. Those are hierarchically substitutable in the sense that each reserve type can be used to meet demand for the subsequent types. We analyze alternative auction designs that recognize that substitutability.

SA29.3 Pay-as-Bid Auctions for Ancillary Services

We use an equilibrium bid function approach to analyze the implications of a pay-as-bid auction for hierarchically substitutable ancillary services. Bids specify reserve type and price per MW. The system operator selects bids so as to meet realized demand for all ancillary services at minimum total cost and each selected bid is paid its bid price.

SA29.4 Solving for Equilibrium Prices in Security-Constrained, Unit Commitment Auction Markets for Electric Power using MIP
• William R. Stewart; College of William & Mary, Sch. of Bus., PO Box 8795, Williamsburg, VA 23187-8795; william.stewart@business.wm.edu
• Benjamin F. Hobbs; JHU, 313 Ames Hall, DOGEE, 3400 North Charles St., Baltimore, MD 21218; bhobbs@jhu.edu
• Michael H. Rothkopf; Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854-8003; rothkopf@rutcor.rutgers.edu
• Udi Helman; Federal Energy Regulatory Commission, 888 First St. NE, Rm. 61-58, Washington, DC 20426;
• Richard P. O'Neill; Federal Energy Regulatory Commission, Office of Economic Policy, 888 First St. NE, Washington, DC 20426; richard.oneill@ferc.fed.us
• Paul M. Sotkiewicz; Federal Energy Regulatory Commission, 888 First St. NE, Rm. 61-58, Washington, DC 20426; sotkiepm@dale.cba.ufl.edu

We present a comprehensive formulation of the security-constrained unit commitment auction problem for electricity markets. We consider several examples of such auction problems and solve them using MIP. We show how the solution of the MIP can be used to generate prices and quantities that clear the market.

# Connecting Modeling Languages with other Systems

Session: SA30
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Robert Fourer
Chair Address: Northwestern University, Dept. of IEMS, 2145 Sheridan Rd., Tech. Inst., Evanston, IL 60208
Chair E-mail: 4er@iems.nwu.edu
Chair:
Chair E-mail:

SA30.1 Using Shared Libraries (DLLs) to Extend an Algebraic Modeling Language

One way to add flexibility to an algebraic modeling language is to permit introducing certain kinds of extensions by means of shared or dynamic-link libraries (DLLs). We sketch the uses of sharedlibraries and describes their use with AMPL language to introduce new functions and table handlers.

SA30.2 AIMMS as an Object in Other Applications
• Johannes J. Bisschop; Paragon Decision Technology BV, PO Box 3277, Haarlem, 2001 DG , The Netherlands; j.j.bisschop@paragon.nl
• Marcel Roelofs; Paragon Decision Technology, PO Box 3277, Haarlem, 2001 DG , The Netherlands; marcel.roelofs@paragon.nl

We will focus on the usability of AIMMS models as optimization components in other applications such as spreadsheets and e-commerce packages. Special attention will be paid to an Excel add-in and an intuitive COM object library to interface with languages such as Visual Basic and VBScript.

SA30.3 Hooking Non-Traditional Solvers to an Algebraic Modeling Language
• Robert Fourer; Northwestern University, Dept. of IEMS, 2145 Sheridan Rd., Tech. Inst., Evanston, IL 60208; 4er@iems.nwu.edu
• David M. Gay; Bell Labs, Lucent Technologies, 600 Mountain Ave., Murray Hill, NJ 07974; dmg@research.bell-labs.com

New optimization techniques are forcing algebraic modeling languages to adopt new approaches in communicating problems to solvers. Global optimization and constraint programming solvers may need to see explicit expression trees, while accepting diverse nonsmooth and combinatorial expression forms. We take examples from several solver interfaces to the AMPL modeling language.

SA30.4 Interactive GUI Objects in AIMMS
• Marcel Roelofs; Paragon Decision Technology, PO Box 3277, Haarlem, 2001 DG , The Netherlands; marcel.roelofs@paragon.nl
• Johannes J. Bisschop; Paragon Decision Technology BV, PO Box 3277, Haarlem, 2001 DG , The Netherlands; j.j.bisschop@paragon.nl

In some planning and scheduling applications, the end-user wants to manipulate both the input and output data in a graphical manner. We will focus on 2 of the several interactive GUI objects in AIMMS: the advanced interactive network object for spatial planning and the flexible interactive Gantt chart for scheduling.

# Technologies in Supply Chain Management

Session: SA31
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Ramesh Sharda
Chair Address: Oklahoma State University, 329 Coll. of Bus. Admin., Stillwater, OK 74078
Chair E-mail: sharda@okstate.edu
Chair:
Chair E-mail:

SA31.1 Network Security & the Business Value-Chain: Past, Present & Future
• Susan J. Chinburg; Oklahoma State University, B4-CBA, Stillwater, OK 74078; schinbu@okstate.edu
• Mark Weiser; Oklahoma State University, B4-CBA, Stillwater, OK 74078; weiser@okstate.edu
• Ramesh Sharda; Oklahoma State University, 329 Coll. of Bus. Admin., Stillwater, OK 74078; sharda@okstate.edu

E-commerce is changing the nature of business. Computer networks and security of those networks have become critical business functions. We will review the related network security issues for each part of the business value chain. Network security processes will be reviewed along with recommendations for future research.

SA31.2 Flexible Shop Floor Control Systems
• Baskar Krishnamoorthy; Oklahoma State University, 322 Engineering North, Stillwater, OK 74078; bask@okstate.edu
• Manjunath Kamath; Oklahoma State University, 322 Engineering North, Stillwater, OK 74078; mkamath@okstate.edu

In today's dynamic and competitive environment, frequent reconfigurations are becoming increasingly necessary on the shop floor. Our focus is on the design of flexible shop floor control systems that can accommodate changes to the physical system and control/decision-making logic. Some preliminary work on the development of a generic control architecture will be presented.

SA31.3 A User-Oriented Framework for Process & Performance Modeling of Enterprise Systems
• Manjunath Kamath; Oklahoma State University, 322 Engineering North, Stillwater, OK 74078; mkamath@okstate.edu
• Nikunj Dalal; Oklahoma State University, B4-CBA, Stillwater, OK 74078; nik@okstate.edu
• William Kolarik; Oklahoma State University, 321D Engineering North, Stillwater, OK 74078; kolarik@okstate.edu
• Amy Hing-Ling Lau; Oklahoma State University, B4 CBA, Stillwater, OK 74078; acctahl@okstate.edu

We explore preliminary issues relating to the design, development, and testing of a user-oriented framework for process and performance modeling of enterprise (ERP) systems. The goal of this interdisciplinary project - just funded by NSF - is to integrate specific methods from engineering and business for improving enterprise-wide processes.

SA31.4 Evaluation of Cuts & Formulations of the Quadratic Knapsack Problem

The QKP is known to be much harder than the linear knapsack problem of similar size. It is encountered in important applications, from location to telecommunications and compiler design. We analyze, both analytically and computationally, several integer programming approaches for solving the QKP.

# Trust & Governance in Interorganizational Relations

Session: SA32
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair: Bill McEvily
Chair Address: Carnegie Mellon University, GSIA, Pittsburgh, PA 15213
Chair E-mail: bmcevily@andrew.cmu.edu
Chair:
Chair E-mail:

SA32.1 Trust Asymmetries in Interorganizational Dyads: Exploring Causes & Consequences

We challenge the assumption that exchange partners hold a symmetric view of their interorganizational relationship. We explore whether trust is mutual and the extent to which the determinants and consequences of trust vary across the dyad. Data from both sides of 99 buyer-supplier dyads are used to test predictions.

SA32.2 The Interplay of Interpersonal & Interorganizational Trust in Biotechnology Alliances
• Akbar Zaheer; University of Minnesota, Carlson Sch. of Mgmt. 3-370, 321 19th Ave. South, Minneapolis, MN 55455; azaheer@csom.umn.edu
• Shawn Lopstrom; University of Maryland, Mgmt. & Org. Dept., Smith Sch., 3340 Van Munching Hall, College Park, MD 20742; slofstro@rhsmith.umd.edu
• Varghese George; Rutgers University, Fac. of Mgmt., Org. Mgmt. Dept, 111 Washington St., MEC 308, Newark, NJ 07102-3027; vgeorge@andromeda.rutgers.edu

We examine interpersonal trust across organizational boundaries and the relationship between interpersonal and interorganizational trust, based on interviews from both sides of 6 biotechnology alliances. We find that interorganizational trust is not merely an aggregation of interpersonal trust and the 2 forms of trust are sometimes driven in opposite directions.

SA32.3 Substitutes or Complements? Exploring the Relationship between Formal Contracts & Relational Governance
• Laura Poppo; Virginia Tech, Dept. of Mgmt., Pamplin 2007 (0233), Blacksburg, VA 24061; lpoppo@vt.edu
• Todd R. Zenger; Washington University, Olin Sch. of Bus., CB 1133, 1 Brookings Dr., St. Louis, MO 63130; zenger@olin.wustl.edu

Trust and relational governance are commonly viewed as substitutes for complex, formal contracts in exchange relations. Indeed, some argue that formal contracts promote opportunistic behavior that destroys trust. We develop and test an alternative perspective that formal contracts and relational governance function as complements. Using data from a sample of information service exchanges, we find empirical support for this complementary relationship.

SA32.4 The Impact of Network Structure on Business Survival after Customers Fail: A Routine-Based View
• Anand Swaminathan; University of California, Grad. Sch. of Mgmt., 1 Shields Ave., Davis, CA 95616-8609; aswaminathan@ucdavis.edu
• Glenn Hoetker; University of Michigan, Bus. School, Ann Arbor, MI 48109-1234; ghoetker@umich.edu
• Will Mitchell; University of Michigan, Bus. School, 701 Tappan St., Ann Arbor, MI 48109-1234; wmitchel@umich.edu

How does the structure of buyer-supplier networks affect supplier survival following customer failure? We test propositions concerning auto sector customer size and status, relationship duration and supplier autonomy. We compare effects for modular and architectural components and for customer acquisition and failure. The study follows a routine-based theory of strategy.

# Flexible Manufacturing Systems

Session: SA33
Date/Time: Sunday 08:30-10:00
Type: Contributed
Track:
Cluster:
Room:
Chair: Sheo G. Misra
Chair Address: Misra Associates, 5 Catalpa St., Morgantown, WV 26505-3677
Chair E-mail:
Chair:
Chair E-mail:

SA33.1 An Algorithm Combining Computer Simulation & Intelligent Searching Method for the Machine Layout in an FMS
• Joon-Mook Lim; Waseda University, Dept. of IMSE, 3-4-1 Shinjuku-ku, Tokyo, 169-8555 , Japan; iac99082@mn.waseda.ac.jp
• Kazuho K. Yoshimoto; Waseda University, Dept. of IMSE, 3-4-1 Okubo Shinjuku-ku, Tokyo, 169-8555 , Japan; kazuho@yoshi.mgmt.waseda.ac.jp
• Jae Kook Lim; Waseda University, Inst. of Asia-Pacific Studies, Sodai-Nishiwaseda Bldg. 6F, Tokyo, 169-0051 , Japan; jklim@mn.waseda.ac.jp

We deal with the layout problem of an integrated automated manufacturing system which has a number of machine centers placed along a built-in automated storage/retrieval system. Recognizing the limits of existing approaches, we propose a new algorithm combining the computer simulation and intelligent searching method.

SA33.2 withdrawn - author request of 10/16
• Ranga V. Ramasesh; Texas Christian University, Neeley Sch. of Bus., Box 32868, Fort Worth, TX 76129; r.ramasesh@tcu.edu
• Shailesh S. Kulkarni; University of North Texas, BCIS Dept., Box 305249, Denton, TX 76203-5249; kulkarni@unt.edu
• Maliyakal D. Jayakumar; University of North Texas, BCIS Dept., Box 305249, Denton, TX 76203-5249; jayakuma@unt.edu

SA33.3 Economic & Process Models for Manufacturing Operations
• Sheo G. Misra; Misra Associates, 5 Catalpa St., Morgantown, WV 26505-3677;

Unit manufacturing operations are at the heart of every manufacturing system. Economic and process models are essential tools in designing, developing, planning, optimizing and controlling manufacturing operations and systems. The status of the development and application in these models is presented along with real world problems and challenges that influence their use.

# Decision Analysis I

Session: SA34
Date/Time: Sunday 08:30-10:00
Type: Contributed
Track:
Cluster:
Room:
Chair: Alan J. Brothers
Chair Address: Battelle PNNL, PO Box 999 K8-03, Richland, WA 99352
Chair E-mail: alan.brothers@pnl.gov
Chair:
Chair E-mail:

SA34.1 A Fuzzy Logic Approach to Buffer Control in Asynchronized Transfer Mode Networks
• Constance A. Lightner; Fayetteville State University, 2529-5 Buffalo Church Rd., Sanford, NC 27330; clightner@uncfsu.edu
• Shu-Cherng Fang; North Carolina State University, OR Program, Raleigh, NC 27695-7906; fang@eos.ncsu.edu
• Henry L. Nuttle; North Carolina State University, OR Program, PO Box 7906, Raleigh, NC 27695-7906; nuttle@eos.ncsu.edu

A fuzzy buffer controller is developed to minimize cell loss in ATM switches. The proposed fuzzy controller sequences cells in buffers based upon the corresponding priority class, end-to-end delay parameter and congestion status. Then cells are scheduled and/or discarded according to the sequenced positions.

SA34.2 A Metamodel Approach to Sensitivity Analysis of Capital Investment Decisions
• Ravipim Chaveesuk; University of Pittsburgh, 207 Dunstan Hall, Auburn, AL 36849; racst31@pitt.edu
• Alice E. Smith; Auburn University, Dept. of ISE, 207 Dunstan Hall, Auburn, AL 36849; aesmith@eng.auburn.edu

A metamodel approach and design of experiment can be used to understand the effect of uncertain input factors on investment decisions. Three types of metamodels are investigated for performing sensitivity analysis of a capital investment: polynomial regression, dual kriging and neural networks. The performance of each type of metamodel is illustrated through a case study.

SA34.3 Decision Support for Optimizing Health Care Technology Assessment
• Elliot B. Sloane; Villanova University, Coll. of Commerce & Finance, 800 Lancaster Ave., Villanova, PA 19085; ebsloane@villanova.edu

Medical technology decisions have become very complex, pitting devices, surgical techniques, pharmaceutical compounds and genetically-engineered alternatives against each other. One of the world's leading research laboratories, ECRI, successfully pioneered 2 analytical techniques to identify optimal resource allocation strategies for government and health care providers...

SA34.4 A Database for Multi-Attribute Evaluation

We describe a database for evaluating whether to extend a vendor's multi-billion dollar government contract. The database contained ratings on an objectives hierarchy and documented their basis. It facilitated systematic group evaluation and basis development on limited access computer share drives. It also helped coordinate the evidential basis with the evaluation.

# Forecasting I

Session: SA35
Date/Time: Sunday 08:30-10:00
Type: Contributed
Track:
Cluster:
Room:
Chair: J. Scott Armstrong
Chair E-mail: armstrong@wharton.upenn.edu,, http://hops.wharton.upenn.edu/forecast
Chair:
Chair E-mail:

SA35.1 Forecast Competition Results in Predicting Monthly Natural Gas Sales in Ohio
• Hisham M. Choueiki; PUCO, 180 East Broad St., 3rd Floor, Columbus, OH 43215; puco_choueik@pucvms.a1.ohio.gov
• David L. Wang; PUCO, 180 East Broad St., 3rd Floor, Columbus, OH 43215;

We investigate the accuracy of our previously built neural network forecasting model with the 'more traditional' ARIMA time series model in predicting the monthly natural gas sales for each of the 4 largest LDCs in Ohio. Our results indicate the superiority of the neural network model over the ARIMA model.

SA35.2 Updating Arima Predictions for Temporal Aggregates
• Yue Fang; University of Oregon, Coll. of Bus. 1208, Eugene, OR 97403; yfang@darkwing.uoregon.edu
• Sergio Koreisha; University of Oregon, Coll. of Bus. 1208, Eugene, OR 97403;

We develop and extend previous investigations on the temporal aggregation of ARMA predictions. Given a basic ARMA model for disaggregated data, 2 set of predictors may be constructed for future temporal aggregates: predictions based on models utilizing aggregated data or on models constructed from disaggregated data for which forecasts are updated as soon as the new information becomes available...

SA35.3 Forecasting Financial Statements
• Anibal C. Irarrazabal; Universidad Catolica de Chile, Vicuna MacKenna 4860, Santiago, , Chile; airarraz@ing.puc.cl

Most financial statements analysis tasks are undertaken with a forward-looking decision in mind. The best way to forecast future performance is to do it comprehensively, by forecasting not only earnings, but also cash flows and balance sheet. Some cases will be presented.

SA35.4 Principles for Forecasting: What We Know & What We Know We Don't Know

Despite much research on forecasting since 1960, advances in forecasting practice are modest. There are 3 reasons: most research is irrelevant to applications, what is relevant has been difficult to use and what has been easily available has been ignored. I provide a DSS to address these issues.

# Inventory Management I

Session: SA36
Date/Time: Sunday 08:30-10:00
Type: Contributed
Track:
Cluster:
Room:
Chair: Arnab Bisi
Chair Address: University of British Columbia, Fac. of Commerce, 2053 Main Mall, Vancouver, BC, V6T 1Z2 , Canada
Chair E-mail: abisi@coe.ubc.ca
Chair:
Chair E-mail:

SA36.1 A Mathematical Evaluation of the Negative Binomial Approximation in a Multi-Echelon Inventory Model
• Adriano O. Solis; University of Texas, Dept. of Info & Dec. Sci., Coll. of Bus. Admin., El Paso, TX 79968-0544; ronnisolis@aol.com
• Charles P. Schmidt; University of Alabama, Dept. of MS & Statistics, Coll. of Comm. & Bus. Admin., Tuscaloosa, AL 35487-0226; cschmidt@cba.ua.edu
• Michael D. Conerly; University of Alabama, Dept. of MS & Statistics, Coll. of Comm. & Bus. Admin., Tuscaloosa, AL 35487-0226; mconerly@cba.ua.edu

Graves (1985), Lee & Moinzadeh (1987) and Graves (1996) provided computational evidence in support of negative binomial approximations to the distributions of certain random variables arising in 3 different multi-echelon inventory models. For the latest model (Graves, 1996), we provide the first mathematical evaluation of the effectiveness of the approximation.

SA36.2 Approximate Optimization of a Two-Level Inventory System

We consider a 2-echelon inventory system with a central warehouse and a number of retailers facing stochastic demand. All installations apply continuous review installation stock (R, Q) policies. There are linear holding and backorder costs. We present a simple technique for approximate optimization of the reorder points.

SA36.3 Optimal Payment Time for a Retailer with Allowed Shortage & Delay in Payment
• A. A. M. Jamal; Southeastern Louisiana University, Dept. of Mgmt., 1104 Rue Cannes, Hammond, LA 70402; ajamal@selu.edu
• Bhaba Sarker; Louisiana State University, Dept. of IMSE, Baton Rouge, LA 70803-6409;
• Shaojun Wang; Louisiana State University, Dept. of IMSE, Baton Rouge, LA 70803-6409; swang2@univx1.sncc.lsu.edu

We develop a model for optimal cycle and payment time for a retailer in deteriorating-item inventory problems where the wholesaler allows a credit period to the retailer for payment without penalty. The model includes planned shortages. The system is modeled for cost minimization to determine the optimal payment under various system parameters. The model is solved by using a iterative procedure.

SA36.4 withdrawn - author request of 10/6
• Eric C. Jackson; Michigan State University, Eli Broad Coll. of Bus., East Lansing, MI 48824; jacks351@pilot.msu.edu
• Ram Narasimhan; Michigan State University, Eli Broad Coll. of Bus., East Lansing, MI 48824; narasimh@msu.edu
• David Mendez; University of Michigan, Health Mgmt. & Policy, Ann Arbor, MI 48109; dmendez@umich.edu

SA36.5 The Role of Information in Supply Chains with Censored Newsvendors
• Arnab Bisi; University of British Columbia, Fac. of Commerce, 2053 Main Mall, Vancouver, BC, V6T 1Z2 , Canada; abisi@coe.ubc.ca
• Martin L. Puterman; University of British Columbia, Ctr. for Op. Excellence, 2053 Main Hall, Vancouver, BC, V6T 1Z2 , Canada; marty@coe.ubc.ca
• Xiaomei Ding; University of British Columbia, Fac. of Commerce, 2053 Main Mall, Vancouver, BC, V6T 1Z2 , Canada; xiaomei_ding@i2.com

We consider 2-echelon supply chains with 1 supplier and 2 retailers. Retailers are censored newsvendors facing iid demand distributions involving unknown parameters. Using Bayesian MDP formulation, we investigate how a supplier can use combined information obtained from retailers' sales data to increase channel profits in a wholesale pricing contract environment.

# Optimization Techniques I

Session: SA37
Date/Time: Sunday 08:30-10:00
Type: Contributed
Track:
Cluster:
Room:
Chair: Illya V. Hicks
Chair Address: Texas A&M University, Dept. of IE, 241 Zachry Engineering Ctr., College Station, TX 77843
Chair E-mail: ivhicks@caam.rice.edu
Chair:
Chair E-mail:

SA37.1 Application of Tabu Search to a Special Class of Multicommodity Distribution Systems
• Siavash Pasalar; Sharif University of Technology, Dept. of IE, Azadi Ave., Tehran, , Iran; pasalar@me.umn.edu
• Kourosh Eshghi; Sharif University of Technology, IE Dept., Azadi Ave., Tehran, , Iran; eshghi@sina.sharif.ac.ir

In order to have a better distribution system in a multi-plant company, we have formulated the problem as a multi-stage multi-commodity network flow model. For solving the relatively large obtained model we have developed a special version of tabu search technique. The details of the approach and computational results are reported.

SA37.2 Studying Neighborhood Functions for Discrete Optimization Problems

The effectiveness of local search algorithms on discrete optimization problems is influenced by the choice of neighborhood function. Closed form expressions and formulas are given for the number of neighborhood functions, of a particular size, with no local minima that are not global minima.

SA37.3 A Modified Benders Approach as Lower Bound for Manufacturing Supply Chains Cost

A modified Benders approach is presented to obtain a lower bound to the integrated cost of manufacturing supply chains obtained from a network of possible options. Costs considered are setup, production, holding and backorder at the manufacturing locations and transportation and selection cost at the links between them. The integrated cost is obtained...

SA37.4 Got Minor?
• Illya V. Hicks; Texas A&M University, Dept. of IE, 241 Zachry Engineering Ctr., College Station, TX 77843; ivhicks@caam.rice.edu

Branch decompositions and branchwidth were introduced by Robertson & Seymour. We present an algorithm to test for minor containment in graphs using branch decompositions.

# NLP I

Session: SA38
Date/Time: Sunday 08:30-10:00
Type: Contributed
Track:
Cluster:
Room:
Chair: Kurt M. Bretthauer
Chair Address: Indiana University, Dept. of Op. & Dec. Tech., Kelley Sch. of Bus., Bloomington, IN 47405
Chair E-mail: kbrettha@indiana.edu
Chair:
Chair E-mail:

SA38.1 withdrawn - author request of 11/6
• Ming Y. Wang; Chinese Academy of Sciences, PO Box 2719, Beijing, 100080 , China; wmy@lsec.cc.ac.cn
• Ya X. Yuan; Chinese Academy of Sciences, PO Box 2719, Beijing, 100080 , China; yyx@lsec.cc.ac.cn

SA38.2 An Analytic Center Cutting Plane Method for the Convex Semi-Definite Feasibility Problem

A convex semi-definite feasibility problem is to find a point in a bounded convex set with a non-empty interior. This set contains a full dimensional ball and is contained in a compact convex set defined by convex matrix inequalities, called the set of localization. The semi-definite feasibility problem arises in many semi-definite relaxations of combinatorial optimization...

SA38.3 withdrawn - author request of 10/27
• Garud Iyengar; Columbia University, Mudd Bldg., Rm. 314, 500 West 120th St., New York, NY 10027; garud@ieor.columbia.edu
• Mehmet Tolga Cezik; Columbia University, Mudd Bldg., Rm. 318, 500 West 120th St., New York, NY 10027; cezik@ieor.columbia.edu

SA38.4 The Nonlinear Resource Allocation Problem with Block Diagonal Constraints
• Kurt M. Bretthauer; Indiana University, Dept. of Op. & Dec. Tech., Kelley Sch. of Bus., Bloomington, IN 47405; kbrettha@indiana.edu
• Bala Shetty; Texas A&M University, Dept. of Info. & Op. Mgmt., Mays Coll. of Bus., College Station, TX 77843; bshetty@cgsb.tamu.edu
• Siddhartha Syam; Marquette University, Dept. of Mgmt., Coll. of Bus., Milwaukee, WI 53201; syam@mail.busadm.mu.edu

We present algorithms for solving a class of nonlinear resource allocation problems (or, nonlinear knapsack problems) that include additional specially structured constraints. Both continuous and integer variable versions of the problem are considered. Computational testing of the algorithms will be reported.

# Marketing I

Session: SA39
Date/Time: Sunday 08:30-10:00
Type: Contributed
Track:
Cluster:
Room:
Chair: Chaim M. Ehrman
Chair Address: Loyola University, 820 North Michigan Ave., Marketing Dept., Chicago, IL 60611
Chair E-mail: cehrman@luc.edu
Chair:
Chair E-mail:

SA39.1 Using Within- & Across-Market Heterogeneity for Optimal Allocation of Promotional Budgets
• Vishal Singh; Northwestern University, Kellogg Grad. Sch. of Mgmt., 2001 Sheridan Rd., Evanston, IL 60208; v-singh1@nwu.edu
• Sachin Gupta; Cornell University, Johnson Grad. Sch. of Mgmt., Sage Hall, Ithaca, NY 14853; s-gupta7@nwu.edu,, sg248@cornell.edu

We develop a random coefficient logit model to explore differences in price sensitivity and effectiveness of promotional tools across markets. Using household panel data from 50 major markets in the US, we decompose the total heterogeneity in parameters into household differences and systematic regional differences. We utilize demographic and lifestyle variables for individual households to explain the unobserved heterogeneity...

SA39.2 no show
• Koushiki Choudhury; Indian Institute of Management, 108 Maniktala Main Rd. SINP, Block 2, Flat 14, Calcutta West Bengal, 700054 , India; koushiki@hotmail.com
• Avinandan Mukherjee; Indian Institute of Management, Alipore Post Office, Box 16757, Calcutta, 700027 , India;

SA39.3 Optimization Model for Media Selection & Planning

We present an optimization model for media selection and planning. Additivity and separable weights properties will be presented. An application in e-commerce will be discussed.

SA39.4 On Streamlining Information Search in Marketing Journals
• Chaim M. Ehrman; Loyola University, 820 North Michigan Ave., Marketing Dept., Chicago, IL 60611; cehrman@luc.edu

When one conducts an information search in the marketing area, one typically finds the discipline to be focused along diversified areas. Research papers can be along the following concentrations: psychology, sociology, anthropology, management, probability and statistics, MS, information science, etc. In order to enhance the efficiency of information search, 47 marketing journals were rated on their ability to disseminate new information to the discipline...

# Queueing Systems

Session: SA40
Date/Time: Sunday 08:30-10:00
Type: Contributed
Track:
Cluster:
Room:
Chair: Greg Miller
Chair Address: Stephen F. Austin State University, Box 13040 -SFA Station, MNB 312, Nacogdoches, TX 75962
Chair E-mail: miller@math.sfasu.edu
Chair:
Chair E-mail:

SA40.1 A Semi-Markov Decision Model with Partial Information for a System with M Parallel Queues & One Server
• Rita M. Rodrigues; Brazilian Institute for Space Research, Av. dos Astronautas, 1758 Jardim da Granja, Sao Jose dos Campos, SP 12201-970 , Brazil; rita@lac.inpe.br
• Solon V. Carvalho; Brazilian Institute for Space Research, Av. dos Astronautas, 1758 Jardim da Granja, Sao Jose dos Campos, SP 12201-970 , Brazil; solon@lac.inpe.br

We want to minimize the long-run expected average cost of a system with M parallel and independent queues and only one server. Customers arrive in each queue according to a Poisson process and the service times have general distribution. This system is modeled as a semi-Markov decision process with partial information.

SA40.2 Optimal Threshold Policies in a Workload Model with a Variable Number of Service Phases per Job

We consider a basic model for 2 essential on-line decisions that must be taken in workload models. The first is the decision to either continue or abort the service of a job. The second concerns the decision to either accept or reject new jobs. We show that under certain regularity conditions, there exist optimal threshold policies for these 2 decisions.

SA40.3 A Queueing Network Model for Capacity Planning of Prisons in the Netherlands
• Rommert Dekker; Erasmus University Rotterdam, Burg. Oudlaan 50, Rotterdam, 3062PA , The Netherlands; rdekker@few.eur.nl,, http://www.few.eur.nl/few/people/rdekker
• Ad Ridder; Free University Amsterdam, de Boelelaan 1010, Amsterdam, , The Netherlands;
• R. Korporaal; Erasmus University Rotterdam, Burg. Oudlaan 50, Rotterdam, 3062PA , The Netherlands;

We describe a queueing network with blocking after service that isembedded in a DSS to support capacity planning of prisons. In particular, we predict the probability that a criminal is sent home because of a shortage of cells.

SA40.4 Queue Network Modeling of Job Shops for Workload Control

Recently, manufacturing strategies have focused on speed of response to customer as much as cost and quality for competitive advantage. Workload control is one such strategy. Its basis is an extension of Little's formula. Exact and approximate open queueing network models for general job shops are presented.

SA40.5 Estimation of the Coefficient of Variation for Unobservable Times in the M/G/1 Queue
• Greg Miller; Stephen F. Austin State University, Box 13040 -SFA Station, MNB 312, Nacogdoches, TX 75962; miller@math.sfasu.edu
• U. Narayan Bhat; SMU, Box 750240, Dallas, TX 75275; nbhat@mail.smu.edu

An estimator of the service-time coefficient of variation is presented for the purpose of discriminating between families of distributions useful in the M/G/1 queue. Assuming the service time random variable is unobservable, we base all of our inference on data collected from the imbedded Markov chain in the M/G/1 system.

# DEA Applications in Accounting

Session: SA41
Date/Time: Sunday 08:30-10:00
Sponsor: Accounting, Auditing & Tax Section
Track:
Cluster:
Room:
Chair: Rajiv Banker
Chair Address: University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO43, Richardson, TX 75083-0688
Chair E-mail: rbanker@utdallas.edu
Chair:
Chair E-mail:

SA41.1 Fundamental Analysis of Stocks via Two-Stage Data Envelopment Analysis

Fundamental analysis of stocks links financial data to firm value in 2 consecutive steps, a predictive information link (current financial data - future earnings) and a valuation link (future earnings - firm value), with special emphasis on these market valuations.

SA41.2 Evaluating Super-Efficiency Procedures in Data Envelopment Analysis for Outlier Identification & Ranking Efficient Units
• Hsihui Chang; University of Texas at Dallas, Sch. of Mgmt., MS JO43, Richardson, TX 75083-0688; swchang@utdallas.edu
• Rajiv Banker; University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO43, Richardson, TX 75083-0688; rbanker@utdallas.edu

Monte Carlo simulation experiments are to evaluate the relative performance of the super-efficiency procedures for outlier identification and ranking efficient units. We find that the super-efficiency model does not perform well for ranking efficient units. However, it performs satisfactorily for outlier identification when some observations may be contaminated with random noise.

SA41.3 Predictive Ability of Data Envelopment Analysis Efficiency Scores
• Raj Mashruwala; University of Texas at Dallas, Sch. of Mgmt., MS JO43, Richardson, TX 75083-0688; rajmash@utdallas.edu
• Rajiv Banker; University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO43, Richardson, TX 75083-0688; rbanker@utdallas.edu
• Constantine Konstant; University of Texas at Dallas, Sch. of Mgmt., MS JO43, Richardson, TX 75083-0688; konstans@utdallas.edu

An analytical model demonstrates the incremental information content of DEA efficiency scores in predicting future profits, given current profits. Empirical analysis of data from a chain of retail stores provides evidence in support of the predictive ability of DEA efficiency scores.

# Firm Capabilities, Firm Evolution & Industry Dynamics

Session: SA42
Date/Time: Sunday 08:30-10:00
Track:
Cluster:
Room:
Chair Address: Santa Clara University, Dept. of Org. Analysis & Mgmt., Leavey Sch. of Bus., Santa Clara, CA 95053
Chair:
Chair E-mail:

SA42.1 Managerial Time Horizons & New Product Introduction

No abstract supplied.

SA42.2 Entrant & Incumbent Adjustment to Deregulation
• Tammy L. Madsen; Santa Clara University, Dept. of Org. Analysis & Mgmt., Leavey Sch. of Bus., Santa Clara, CA 95053; tmadsen@scu.edu
• Gordon B. Walker; SMU, Cox Sch. of Busl, PO Box 750333, Dallas, TX 75275-0333; gwalker@mail.cox.smu.edu

We compare incumbent and entrant adjustment to deregulation as a fundamental institutional change in the trucking industry, 1980-1993. Using an evolutionary model of firm growth, findings illustrate that entrants adjust faster to the new regime and that their production function has a more predictable pattern of growth compared to incumbent firms...

SA42.3 Exploiting Exploration
• Anne Marie Knott; University of Pennsylvania, The Wharton Sch., 2021 Steinberg Hall, Philadelphia, PA 19104-6370;

One of the more persistent themes in the management literature is the tension between success in a competitive environment vs. survival in a changing environment. Success in the competitive environment involves exploitation of a firm's existing capabilities (March, 1991 Levinthal & March, 1993). Survival in a dynamic environment involves exploration for new capabilities (March, 1991Levinthal & March, 1993)...

# Risk Management I

Session: SA43
Date/Time: Sunday 08:30-10:00
Type: Contributed
Track:
Cluster:
Room:
Chair: Arun Verma
Chair Address: Cornell University/FISC, 634 Rhodes Hall, Ithaca, NY 14850
Chair:
Chair E-mail:

SA43.1 Pricing the Credit Risk with Regime Switch Models

The credit risk is the risk that an obligation will not be paid and a loss will result. Using Markov regime switch processes, we generalize the previous researches on credit risks by incorporating both transition risks and short spread risks. The necessity of this generalization is also justified.

SA43.2 The Effects of Financial Liberalization on the Tunisian Banking Industry: A Non-Parametric Approach
• Moez Hababou; York University, Schulich Sch. of Bus., Rm. 344, 4700 Keele St., North York, Ontario, M3J 1P3 , Canada; mhababou@ssb.yorku.ca
• Wade D. Cook; York University, Schulich Sch. of Bus., Rm. 340, 4700 Keele St., NOrth York, Ontario, M6J 3P3 , Canada; wcook@ssb.yorku.ca
• Gordon Roberts; York University, Schulich Sch. of Bus., Rm. 340, 4700 Keele St., NOrth York, Ontario, M6J 3P3 , Canada;

We examine the impacts of similar reforms on the efficiency of the banking system in Tunisia, a country whose economy has been reshaped by IMF/World Bank prescribed economic adjustment plans since 1987. Using various DEA models and panel data covering 1992-1997, we evaluate the individual effects of each component of the reforms on the banking industry overall.

SA43.3 withdrawn - author request of 9/18

SA43.4 Dynamic Hedging using a Reconstructed Local Volatility Function

We compare the dynamic hedging performance of the deterministic local volatility function method with the usual constant and implied volatility approach. As is well know, implied volatilities of stock option prices show a smile behavior; this means that the constant volatility Black-Scholes model does not capture the dynamics of the underlying price correctly...

# Practitioners' Award Competition

Session: SB01
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Detlof Von Winterfeldt
Chair Address: University of Southern California, Sch. of Policy/Planning/Dev., Los Angeles, CA 90089
Chair E-mail: detlof@aol.com
Chair:
Chair E-mail:

SB01.1 Practioners' Award Competition

Each year, the Decision Analysis Society of INFORMS presents an award for the best application of decision analysis. The finalists for the award will each present a summary of their work with the winner announced in the afternoon awards session.

# New Directions in Behavioral Research

Session: SB02
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Jean L. Kahwajy
Chair Address: International Institute for Management Development, 23 Chemin de Bellerive, PO Box 915, Lausanne, CH-1001 , Switzerland
Chair E-mail: kahwajy@imd.ch
Chair:
Chair E-mail:

SB02.1 Subjective Probability Judgments: Partition Dependence
• Robert T. Clemen; Duke University, Fuqua Sch. of Bus., Box 90120, Durham, NC 27708-0120; clemen@mail.duke.edu
• Craig Fox; Duke University, Fuqua Sch. of Bus., PO Box 90120, Durham, N- 27708-0120; cfox@mail.duke.edu

We present evidence that subjective probabilities depend critically on how the event space is partitioned. Our results suggest that judges compromise between 'ignorance prior' probability (equal mass across partitions) and probability based on support for the target event (as in support theory). We discuss implications for probability elicitation in practice.

SB02.2 Measuring Financial Investor's Risk Aversion

We seek to devise a method for measuring an investor's risk aversion that fulfills 2 requirements: it asks preference questions meaningful to the individual and the resulting measure will translate into financial portfolio selection in a theoretically sound fashion.

SB02.3 Consulting Experiences in Decision Analysis
• Elizabeth Ewing; Strategic Decisions Group, 2440 Sand Hill Rd., Menlo Park, CA 94025;

The hard part of decision consulting is integrating ideas, analysis and implementation with diverse personalities and competing interests. Some tricks of the trade have been discovered and will be discussed.

SB02.4 How to Hear & How to be Heard
• Jean L. Kahwajy; International Institute for Management Development, 23 Chemin de Bellerive, PO Box 915, Lausanne, CH-1001 , Switzerland; kahwajy@imd.ch

The hard part of decision making is interpersonal communications. This research concerns the role that individuals play in reversing negative situations. It offers a radically different approach to negotiations and a productive strategy for being heard and for overcoming prevailing erroneous or negative expectations.

# System Dynamics in Operations Management

Session: SB03
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Edward G. Anderson
Chair Address: University of Texas, Dept. of Mgmt., CBA 4.202, Austin, TX 78712
Chair E-mail: edward.anderson@bus.utexas.edu
Chair:
Chair E-mail:

SB03.1 Feedback Dominance in the Service Profit Chain
• Rogelio Oliva; Harvard Business School, Soldiers Field Rd., Bakery Library Rm. 182, Boston, MA 02163; roliva@hbs.edu

The service profit chain is a widely used conceptual framework linking employee satisfaction, customer satisfaction and loyalty and financial performance. The framework relies on a series of reinforcing mechanisms. A system dynamics model is used to determine the dominant feedback mechanisms under different scenarios and derive policies to improve performance.

SB03.2 Information Relvance & Rigor as Antecedents to Strategic Clarity: Using SD to Test SAP's Strategic Enterprise Management Module with BS
• James L. Ritchie-Dunham; University of Texas, Dept. of MSIS, 3615 Aspen Creek Parkway, Austin, TX 78749; jimrd@sdsg.com

Decision makers often lack clarity in how to satisfy their multiple stakeholders, because they use improperly filtered and irrelevant information for decisionmaking. ERPs claim to fix this managerial uncertainty in strategic decisionmaking. This research tests that claim and explores an explanation of the process

SB03.3 Using Real Options to Value Flexibility in Product Development Projects

Product development projects suffer regularly from unexpected changes in requirements which induce significant rework. Flexible product architectures can reduce the resulting rework, but only at a price. We use real options theory in a system dynamics model to determine how much a firm should pay for this flexibility

# Logistics Modeling & Analysis

Session: SB04
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Ananth V. Iyer
Chair Address: Purdue University, Krannert Sch. of Mgmt., West Lafayette, IN 47907-1310
Chair E-mail: aiyer@mgmt.purdue.edu
Chair:
Chair E-mail:

SB04.1 Lot-Sizing under Postponement & Downward Substitution for a Perishable Product
• Jayashankar Swaminathan; University of California, Berkeley; msj@haas.berkeley.edu
• Simge Kukukyavuz; University of California, Dept. of IEOR, Berkeley, CA 94720-1900;

We study a lot-sizing problem from the bio-technology industry with 3 interesting characteristics related to inventory management: perishability, postponement and substitution. The product considered is a reagent for detecting DNA sequences. We find the optimal lot-size under constant and varying deterministic demands and evaluate the benefits of investing in technology to mitigate the constraints on perishability and conversion.

SB04.2 An Integrated Supply Chain Model with Lead-Time Dependent Demand
• Saibal Ray; University of Waterloo, MS Dept., Waterloo, Ontario, N2L 3G1 , Canada;
• Elizabeth M. Jewkes; University of Waterloo, MS Dept., Waterloo, Ontario, N2L 3G1 , Canada;
• Yigal Gerchak; University of Waterloo, Dept. of MS, Waterloo, Ontario, N2L 3G1 , Canada;

We model an integrated supply chain where customer demand is a function of price and announced lead-time (AL) while price itself depends on AL. Operating costs exhibit economies of scale. The objective is to maximize the firms' profit, taking into consideration revenue and WIP holding, investment and raw material inventory costs.

SB04.3 Selling to a Newsvendor who Incurs Operating Costs
• Richard Cho; University of Waterloo, Dept. of MS, Waterloo, Ontario, N2L 3G1 , Canada;
• Yigal Gerchak; University of Waterloo, Dept. of MS, Waterloo, Ontario, N2L 3G1 , Canada;

The coordination literature models retailers as costless operations, other than costs of mismatches between demand and product availability. Yet retailers incur operating costs and part of the drive for efficiency involve controlling them. We endow Lariviere & Porteus' selling to the newsvendor model with retail operating costs and explore the implications.

SB04.4 Managing Capacity under Mixed Steady & Random Demand
• Ananth V. Iyer; Purdue University, Krannert Sch. of Mgmt., West Lafayette, IN 47907-1310; aiyer@mgmt.purdue.edu
• Apurva Jain; University of Washington, Seattle, WA;

We model the use of capacity when demand is a mixture of steady and random orders. We model the lead time and inventory cost impact of the following approaches pooled capacity, splitting capacity and use of scheduling priorities. We discuss managerial implications of each of these approaches.

# Tutorial: Moment Problems & their Applications in Probability Theory, Finance, Stochastic Networks & Combinatorial Opt.

Session: SB05
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Dimitris Bertsimas
Chair Address: MIT, Sloan Sch. of Mgmt., E53-359, OR Ctr., 50 Memorial Dr., Cambridge, MA 02139
Chair E-mail: dbertsim@mit.edu
Chair:
Chair E-mail:

SB05.1 Tutorial: Moment Problems & their Applications in Probability Theory, Finance, Stochastic Networks & Combinatorial Optimization
• Dimitris Bertsimas; MIT, Sloan Sch. of Mgmt., E53-359, OR Ctr., 50 Memorial Dr., Cambridge, MA 02139; dbertsim@mit.edu

Problems involving moments of random variables arise naturally in many areas of mathematics, economics and OR. How do we obtain optimal bounds on the probability that a random variable belongs in a set, given some of its moments? How do we price financial derivatives without assuming any model for the underlying price dynamics, given only moments of the price of the underlying asset?...

# Sequential & Combined Travel Forecasting Models

Session: SB06
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: David E. Boyce
Chair Address: University of Illinois, Dept. Civil & Materials Eng., 3073 ERF, 842 West Taylor St., Chicago, IL 60607-7023
Chair E-mail: dboyce@uic.edu
Chair:
Chair E-mail:

SB06.1 Propagation of Uncertainty in Multi-Stage Travel Demand Models
• Yong Zhao; University of Texas, Dept. of Civil Eng., ECJ 6.9, C1761, Austin, TX 78712;
• Kara Kockelman; University of Texas, Dept. of Civ. Eng., ECJ 6.9, C1761, Austin, TX 78712; kkockelm@mail.utexas.edu

Large-scale travel demand models are generally estimated sequentially, with results of one model acting as inputs to subsequent models. Typically, only point estimates are passed forward, rather than estimates of variation and covariation. We investigate the nature of uncertainty propagation with typical travel demand models, thereby assessing the impact of the neglect of such variations.

SB06.2 An Origin-Based Approach to Solving a Combined Travel Choice Model
• David E. Boyce; University of Illinois, Dept. Civil & Materials Eng., 3073 ERF, 842 West Taylor St., Chicago, IL 60607-7023; dboyce@uic.edu
• Hillel Bar-Gera; Ben-Gurion University of the Negev, Dept. of IEM, Beer-Sheva, , Israel;

A combined model of origin-destination, mode and route choice for the Chicago region is solved with a new algorithm based on Bar-Gera's origin-based method. Accuracy of the solution and performance of the algorithm are compared to the traditional Evans algorithm for solving this problem.

SB06.3 A New Strategy for Accelerating the Frank-Wolfe Algorithm for the Traffic Assignment Problem
• Der-Horng Lee; Naational University of Singapore, Dept. of Civil Eng., Singapore, 117576 , Singapore; dhl@nus.edu.sg
• Yu Nie; Naational University of Singapore, Dept. of Civil Eng., Singapore, 117576 , Singapore;

A new strategy for accelerating the Frank-Wolfe algorithm is presented. This strategy modifies the search direction by unlimitedly combining previous extreme points. Based on abundant computational results, the proposed modified Frank-Wolf algorithm is appropriate when an accurate solution is needed.

SB06.4 Micro-Assignment of Travel Demand with Activity/Trip Chains
• Ahmded F. Abdelghany; University of Texas, Dept. of Civ. Eng., ECJ 6.2, Austin, TX 78712; afaissal@mail.texas.edu
• Hani S. Mahmassani; University of Texas, Dept. of Civil Eng., ECJ 6.2, Austin, TX 78712-1076; masmah@mail.utexas.edu

A simulation-assignment model of travel demand with activity chains is presented. Experiments are conducted to compare trip-based assignment (with conventional models) to the proposed approach which recognizes trip chains. Pitfalls of inappropriately recognizing trip chains using current practice assignment models are illustrated.

# Skills Assessment for OR/MS

Session: SB07
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Gary W. Shulberg
Chair E-mail: gwshulberg@dscp.dla.mil
Chair:
Chair E-mail:

SB07.1 Genesis of an Assessment of Analytical Skills for a Workforce Development Program
• Allan Rosen; Defense Supply Center Philadelphia, 700 Robbins Ave., Bldg. 3B, #5149, Philadelphia, PA 19111-5096; arosen@dscp.dla.mil

DSCP has initiated a program to define the diversity of business skills required of its current and future workforce, to assess skill levels, and to enhance capabilities. We will discuss the research involved and the obstacles encountered in the creation of this program and their importance to OR/MS educators and practitioners.

SB07.2 Development & Analysis of an Assessment for a Workforce Development Program
• Margaret G. Barton; US Office of Personnel Management, 230 South Dearborn St., DPN 30-3, Chicago, IL 60604; mgbarton@opm.gov
• Patrick J. Sharpe; US Office of Personnel Management, 1900 E St. NW, Rm. 6500, Washington, DC 20415; pjsharpe@opm.gov

A 2-phased process, designed to assess the proficiency of a contracting workforce sample with regard to competencies identified as critical to job performance in a changing workplace, will be presented. Scoring procedures, as well as ways in which OR/MS educators and practitioners can be informed by the results, will be discussed.

SB07.3 Using Webforms to Gather Perceptions of Technical Expertise

Are incoming freshman or seasoned faculty more comfortable using technology? Web-based survey documents are used to gather information about the degree to which School of Business students and faculty have integrated technology in their lives. Initial results, including student and faculty perceptions of each other's technical expertise, are presented.

SB07.4 Using Webforms to Support the Design of an Information Technology Course
• David D. Hott; Florida Institute of Technology, 150 West University Blvd., Melbourne, FL 32901; dhott@raven-villages.net,, dhott@fit.edu
• Judith A. Barlow; Florida Institute of Technology, 150 West University Blvd., Melbourne, FL 32901; jbarlow@fit.edu,, http://winnie.fit.edu/~jbarlow
• Karen Chambliss; Florida Institute of Technology, Sch. of Bus., 150 West University Blvd., Melbourne, FL 32901; kchambl@fit.edu
• David Dorsett; Florida Institute of Technology, Sch. of Bus., 150 West University Blvd., Melbourne, FL 32901;
• Terry Lease; Florida Institute of Technology, Sch. of Bus., 150 West University Blvd., Melbourne, FL 32901;
• Roger Manley; Florida Institute of Technology, Sch. of Bus., 150 West University Blvd., Melbourne, FL 32901;
• Barbara Gugliotta Pierce; Florida Institute of Technology, Sch. of Bus., 150 West University Blvd., Melbourne, FL 32901;
• Michael Slotkin; Florida Institute of Technology, Sch. of Bus., 150 West University Blvd., Melbourne, FL 32901;

New IT tools continue to evolve so rapidly that it is necessary to continuously review which tools to include in a core business curriculum. A collaborative strategy for developing and maintaining a core set of IT skills is presented. Web-based survey documents are used to communicate and document the process.

# Financial Management I

Session: SB08
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Liqun Xu
Chair Address: National University of Singapore, Dept. of Finance & Acct., FBA1 02-03, 15 Law Link, Singapore, 117591 , Singapore
Chair E-mail: fbaxulq@nus.edu.sg
Chair:
Chair E-mail:

SB08.1 Optimal Portfolio of Reconfigurable & Dedicated Machining Systems under Demand Uncertainty
• Wichai Narongwanich; University of Michigan, IOE Dept., 2205 Hubbard St., Apt. 6, Ann Arbor, MI 48105; wichai@engin.umich.edu
• Izak Duenyas; University of Michigan, Kresge Sch. of Bus., Ann Arbor, MI 48105; duenyas@umich.edu
• John R. Birge; Northwestern University, McCormick Sch. of Engineering, 2145 Sheridan Rd., Evanston, IL 60208-3100; jrbirge@nwu.edu

We determine the optimal portfolio of capacity, consisting of DMS capable of producing one product generation and RMS capable of producing several product generations, to satisfy demand over time. The inter-arrival time of 2 consecutive product generations is uncertain and the new product will completely replace the old product. Interesting structural results are provided.

SB08.2 Bank Stock Return Sensitivities to the Long- & Short-Term Interest Rates: A Bivariate GARCH Approach

A bi-variate GARCH methodology is employed to investigate the relative sensitivities of bank stock returns to short- and long-term interest rates and their respective volatilities. Results indicate that the short-term interest rate and its volatility have greater impact on bank portfolio returns and volatility than the long-term rate. Innovations in short- and long-term interest rates are not captured equally by movements in the market index.

SB08.3 Approximate Pricing & Exercising of High-Dimensional American Options: A Duality Approach

We propose an algorithm for pricing high-dimensional American options based on approximate dynamic programming using neural networks to approximate the value function. We also develop upper and lower bounds on option prices which are general enough to be used in conjunction with other approximate methods of pricing American options.

SB08.4 withdrawn - author request of 11/2

# Applications of Revenue Management I

Session: SB09
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Georgia Perakis
Chair Address: MIT, Sloan Sch. of Bus., OR Ctr., 50 Memorial Dr., Rm. E53-359, Cambridge, MA 02139
Chair E-mail: georgiap@mit.edu
Chair:
Chair E-mail:

SB09.1 Pricing Models for the Sea Cargo Industry

The Ocean Shipping Reform Act came into effect on May 1, 1999. The Act allows ocean carriers to enter into confidential contracts with shippers, charging prices that depend not only on the route and commodity, but also on the shipper and the time of year. We discuss some pricing models that carriers can use.

SB09.2 Golf Revenue Management
• Sheryl E. Kimes; Cornell University, Sch. of Hotel Admin., 335A Statler, Ithaca, NY 14853; skimes@twcny.rr.com
• Alfonso Delgado-Muerza; Royal Carribean Cruise Lines;

Golf courses possess the conditions necessary for the application of RM. We present a strategic framework for the application of RM to golf and present the results of an extensive study on the implementation of RM at a university golf course.

SB09.3 Restaurant Revenue Management
• Romy Nakagawa; MIT, OR Ctr., E40-130, 1 Amherst St., Cambridge, MA 02139; romy@mit.edu
• Dimitris Bertsimas; MIT, Sloan Sch. of Mgmt., E53-359, OR Ctr., 50 Memorial Dr., Cambridge, MA 02139; dbertsim@mit.edu

Utilizing integer, approximate dynamic, stochastic programming and stochastic approximation methods, we develop new policies for revenue maximization in a restaurant by determining when, if at all, to seat an incoming party. Computational results show that our methods enhance revenue relative to the industry practice of FCFS by 3-10%.

SB09.4 Revenue Management in the Multi-Family Housing Industry
• Donald M. Davidoff; Talus Solutions, Inc., 4751 Best Rd., Waterstone, Ste. 300, Atlanta, GA 30337; ddavidoff@talussolutions.com
• Nancy M. Pyron; Talus Solutions, Inc., 4751 Best Rd., Waterstone, Ste. 300, Atlanta, GA 30337;

Multi-family housing represents a new frontier for RM. This industry shares many characteristics with traditional RM industries, but presents new challenges, e.g. extremely long lengths of stay and relatively small transaction density. We explore similarities and differences between multi-family housing and traditional RM industries and discuss several of the most pressing challenges that must be met to solve this intriguing challenge.

# Innovations & Applications of Integer Programming

Session: SB10
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: Scheduling
Room:
Chair: Renato E. de Matta
Chair Address: University of Iowa, 108 Pappajohn BAB, Iowa City, IA 52242
Chair E-mail: renato-dematta@uiowa.edu
Chair:
Chair E-mail:

SB10.1 The Capacity Expansion Problem
• Renato E. de Matta; University of Iowa, 108 Pappajohn BAB, Iowa City, IA 52242; renato-dematta@uiowa.edu
• Vernon N. Hsu; George Mason University, Sch. of Mgmt., 4400 University Dr., Fairfax, VA 22030; vhsu@som.gmu.edu

We investigate a dynamic capacity expansion problem with capacity deterioration. We consider situations where the deterioration rate depends on the capacity's acquisition period. In a MIP model, we capture the capacity acquisition and replacement decisions and develop an efficient solution procedure. Computational results are presented.

SB10.2 A Linear Continuous-Time Model in Batch Sizing/Schedulng with a Minimum Number of Events

Let an event represent a time around which at least one product level changes. The resulting event-indexed MILP model requires fewer binary variables than time-indexed models and it handles unequal batch-length, product re-usage, merging and diverging production lines and minimum and maximum batch sizes. We will present some computational results.

SB10.3 Integrated Scheduling of Job Processing & Job Delivery
• Zhi-Long Chen; University of Pennsylvania, Dept. of Systems Eng., 220 South 33rd St., Philadelphia, PA 19104-6315; zlchen@seas.upenn.edu
• Chung-Yee Lee; Texas A&M University, 237 Zachry Engineering Ctr., Dept. of IE, College Station, TX 77843-3131; cylee@acs.tamu.edu

We present models that integrate schedules for processing jobs in job shops and schedules for delivering finished jobs to customers. We will investigate problem complexity and propose algorithms for some cases.

SB10.4 The Work Scheduling Problem with Job Assignment Flexibility

We find the optimal mix of jobs in work schedules for several groups of workers with non-homogeneous skills. A B&P procedure is proposed which minimizes total labor cost. We demonstrate this procedure in developing work schedules for bus drivers.

# Combinatorial Optimization Algorithms in Computer Network Applications

Session: SB11
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: Integer Programming
Room:
Chair: Olivier Goldschmidt
Chair Address: Make Systems, 200 Regency Forest, Cary, NC 27511
Chair E-mail: olivier@makesys.com
Chair:
Chair E-mail:

SB11.1 Local Search with Perturbations for the Prize-Collecting Steiner Tree Problem in Graphs
• Mauricio G. C. Resende; AT&T Labs Research, Information Sci. Research, Florham Park, NJ 07932; mgcr@research.att.com
• S. A. Canuto; Catholic University of Rio de Janeiro, Comp. Sci. Dept., Rua Marques de Sao Vicente 225, Rio de Janeiro, 22453-900 , Brazil; suzana@inf.puc-rio.br
• Celso Ribeiro; Catholic University of Rio de Janeiro, Comp. Sci. Dept., Rua Marques de Sao Vicente 225, Rio de Janeiro, 22453-900 , Brazil; celso@inf.puc-rio.br

Given an undirected graph G with node prizes and edge weights, the prize-collecting Steiner tree problem seeks a subtree of G which minimizes the sum of the weights of its edges plus the prizes of the nodes not spanned. We describe a multi-start local search algorithm (implementation) for this problem.

SB11.2 Combinatorial Approaches to SONET Design Problems
• Eli V. Olinick; SMU, Dept. of Comp. Sci. & Eng., Sch. of Eng. & Applied Sci., Dallas, TX 75275-0122; olinick@seas.smu.edu

We consider problems of minimizing ADM costs in a network of SONET rings. Sites are connected to multiple rings, but traffic between 2 sites cannot be split between rings. We develop heuristics andapproximation algorithms for these difficult design problems by modeling them as edge-partitioning problems on the demand graph.

SB11.3 withdrawn - chair request of 10/19
• Steven T. Cosares; Dowling College, Dept. of Math, Oakdale, NY 11769;

SB11.4 Heuristics for the Exact Topology Network Design Problem

Given a fixed graph topology and a link cost matrix, the exact topology network design problem is to match the nodes of the network with the vertices of the topology to minimize the total cost of the network. We consider heuristics based on assignment problem and on partial solutions of the quadratic assignment problem for solving this problem.

# Technology Supply Chains

Session: SB12
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: Logistics & Supply Chain Management
Room:
Chair: Geoffrey G. Parker
Chair Address: Tulane University, Freeman Sch. of Bus., 7 McAlister Dr., New Orleans, LA 70118
Chair E-mail: geoffrey.parker@tulane.edu
Chair:
Chair E-mail:

SB12.1 Decomposition of Designs within a Technology Supply Chain
• Nitin Joglekar; Boston University, Sch. of Mgmt., 595 Commonwealth Ave., Boston, MA 02215; joglekar@bu.edu
• Charles H. Fine; MIT, Sloan Sch. of Mgmt., Cambridge, MA 02139; charley@mit.edu

We model the payoff from increasing system design performance through task decomposition during a finite market window of opportunity, subject to integration penalty. Optimal outsourcing policy is characterized in terms of design modularity and speeds of performance enhancement, termed clock speed, for the system design firm and its supplier respectively.

SB12.2 Finding a Firm's Equilibrium Clockspeed as a Function of Production Introduction Costs & Industry
• Gilvan C. Souza; University of North Carolina, Kenan-Flagler Bus. Sch., McColl Bldg., CB 3490, Chapel Hill, NC 27599-3490;

We model a market where consumers care about technology and price. Market demand is stationary and split between two firms using a market share attraction model. The problem is formulated as a non-0 sum stochastic game. We find numerical equilibria regarding the firms' clockspeed - pace of new product

SB12.3 From Buyer to Integrator: The Transformation of the Supply Chain Manager in the Vertically Disintegrating Firm

We describe the change in supply-chain management toward supply-chain integration. As large firms disaggregate into linked chains of focused firms specializing in distinct areas, those focused firms paradoxically require more highly skilled generalists, 'supply-chain integrators,' capable of coordinating product development, marketing, production, and logistics within and across organizational boundaries.

# Panel: Research Directions in Manufacturing Logistics

Session: SB13
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: Manufacturing & Logistics
Room:
Chair: S. David Wu
Chair Address: Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015-1582
Chair E-mail: david.wu@lehigh.edu
Chair:
Chair E-mail:

SB13.1 Panel: Research Directions in Manufacturing Logistics
• Morris A. Cohen; University of Pennsylvania, The Wharton School, Dept. of OPIM, 1300 SH-DH, Philadelphia, PA 19104-6366;
• Robin Roundy; Cornell University, Dept. of OR/IE, 216 Rhodes Hall, Ithaca, NY 14853; robin@orie.cornell.edu
• Nicholas G. Hall; Ohio State University, 301 Hagerty Hall, 1775 South College Rd., Columbus, OH 43210-1144; halln@cob.ohio-state.edu

Research in manufacturing logistics expands the traditional scope of production and operations management to include issues such as supply chain coordination, electronic commerce, and the integration of manufacturing and other business functions. This panel will discuss recent and future trends for research in this area.

Session: SB14
Date/Time: Sunday 10:15-11:45
Track:
Cluster: Product Development
Room:
Chair: Vish V. Krishnan
Chair Address: University of Texas, Dept. of Mgmt., CBA 4.202, Austin, TX 78712
Chair E-mail: krishnan@mail.utexas.edu
Chair: Nitin Joglekar
Chair Address: Boston University, Sch. of Mgmt., 595 Commonwealth Ave., Boston, MA 02215
Chair E-mail: joglekar@bu.edu

SB14.1 Panel: Web/IT Driven Product Development

A panel of researchers and practitioners will address the following questions: What are some important research dimensions of web-driven product development? Is this theme a logical extension of research on information-driven design? What are some important trends emerging in practice?

# Tutorial: Coordinating Production, Distribution & Pricing Decisions in the Supply Chain

Session: SB15
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: Quantitative Models in Supply Chain Management
Room:
Chair: David Simchi-Levi
Chair Address: MIT, Dept. of Civil & Environ. Eng., 77 Massachusetts Ave. Rm. 1171, Cambridge, MA 02139
Chair E-mail: dslevi@mit.edu
Chair:
Chair E-mail:

SB15.1 Tutorial: Coordinating Production, Distribution & Pricing Decisions in the Supply Chain
• David Simchi-Levi; MIT, Dept. of Civil & Environ. Eng., 77 Massachusetts Ave. Rm. 1171, Cambridge, MA 02139; dslevi@mit.edu

The influence of the Internet and e-commerce on the economy in general, and supply chain management in particular, has been tremendous. Changes are happening so fast and the scope is breathtaking! General Motors, for instance, has recently announced, (Wall Street Journal 2/22/2000) that the company is completely changing the way it is designing, building, selling and delivering its products...

# Tutorial: Preference Function Modeling

Session: SB16
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: MCDM
Room:
Chair: Jonathan Barzilai
Chair Address: Dalhousie University, Dept. of IE, PO Box 1000, Halifax, Nova Scotia, B3J 2X4 , Canada
Chair E-mail: jonathan.barzilai@dal.ca,, http://www.ScientificDecisions.com
Chair:
Chair E-mail:

SB16.1 Tutorial: Preference Function Modeling

PFM is a new, intuitive, easy-to-use but powerful decision methodology based on a rigorous mathematical theory. This presentation will introduce the basic concepts and theory underlying PFM as well as Trillium/PFM, a software package that implements this methodology.

# Network Flows: Recent Advances in Networks

Session: SB17
Date/Time: Sunday 10:15-11:45
Track:
Cluster: Network Flows
Room:
Chair: Pinar Keskinocak
Chair Address: Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205
Chair E-mail: pinar@isye.gatech.edu
Chair:
Chair E-mail:

SB17.1 Accessing Nearby Copies of Replicated Objects in a Distributed Environment
• Andrea W. Richa; Arizona State University, Dept. of Comp. Sci. & Eng., Box 875406, Tempe, AZ 85287-5406; aricha@asu.edu
• Greg Plaxton; University of Texas, Dept. of Comp. Sci., Taylor Hall 2.124, Austin, TX 78712-1188; plaxton@cs.utexas.edu
• Rajmohan Rajaraman; Northeastern University, Coll. of Comp. Sci., 161 Cullinane Hall, Boston, MA 02115; rraj@ccs.neu.edu

We design a simple, space efficient randomized algorithm for accessing shared replicated objects that tends to satisfy each access request with a nearby copy, ensuring efficient utilization of network resources. We show that under the particular cost model considered, the expected cost of an individual access is asymptotically

• F. Sibel Salman; Carnegie Mellon University, GSIA, 5000 Forbes Ave., Schenley Park, Pittsburgh, PA 15232; fs2c@andrew.cmu.edu

We give an exact solution method for the problem of routing traffic between pairs of nodes of an input graph simultaneously and installing cables on the edges of the graph to accommodate flow at minimum cost. The method relies on a B&B algorithm that solves relaxations (in the form of the multicommodity flow problem)...

SB17.3 Algorithms for Solving Time-Dependent & Dynamic Network Flow Problems
• Sarah Stock Patterson; Duke University, Fuqua Sch. of Bus., Box 90120, Durham, NC 27708-0120; sarah.sp@duke.edu
• Elise D. Miller-Hooks; Pennsylvania State University, Dept. of Civil & Environ. Eng., 212 Sackett Bldg., University Park, PA 16802; edm3@psu.edu

We present algorithms for solving time-dependent and dynamic network flow problems. These problems have the common objective of finding the minimum time paths along which to send the supply between a set of sources and sinks in time-dependent (w.r.t. travel times, arc and node capacities and supply) and dynamic (w.r.t. flow) networks.

SB17.4 A Subadditive Approach to Facets of the Cyclic Group & Knapsack Problem
• Lisa Evans; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA;
• Ellis L. Johnson; Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205; ellis.johnson@isye.gatech.edu
• Ralph E. Gomory; Sloan Foundation;
• Julian Araoz; Universidad Simon Bolivar, Caracas, , Venezuela;

Gomory & Johnson used the subadditive characterization of cyclic group facets to develop a subadditive function approach to integer programming. A master knapsack problem is defined and shown to be closely related to cyclic group polyhedra. Relations between facets and some classes of facets will be shown. Some results from Gomory's shooting experiment will be given.

# Applying Marketing Science to E-Commerce

Session: SB18
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: E-Commerce
Room:
Chair: Lee G. Cooper
Chair Address: UCLA, The Anderson Sch., 110 Westwood Plaza, Ste. B518, Los Angeles, CA 90095-1481
Chair E-mail: lee.cooper@anderson.ucla.edu
Chair:
Chair E-mail:

SB18.1 withdrawn - chair request of 10/17
• Eric T. Bradlow; University of Pennsylvania, Wharton Sch., Ste. 1400 SH-DH, 3620 Locust Walk, Philadelphia, PA 19104; ebradlow@wharton.upenn.edu
• Joel Steckel; NYU, Mgmt. Ed. Ctr., Stern School, 44 West 4th St., New York, NY 10012-1126; jsteckel@stern.nyu.edu

SB18.2 Forecasting Customer Purchases in a Nonstationary Environment
• Bruce G. S. Hardie; London Business School, Sussex Place, Regent's Park, London, NW1 4SA , England, UK;

As e-commerce companies examine the notion of customer lifetime value, the need to forecast purchasing emerges. Traditional approaches assume stationary purchasing (possibly allowing for attrition). However, many new markets are characterized by evolving behavior. We develop a model that captures evolving preferences and attrition, and compare it to benchmark models.

SB18.3 A Segment- & Customer-Oriented Recommendation Engine

The basic e-commerce recommendation engine in use until the beginning of this year has been the collaborative filter - an algorithmic technique that infers the tastes of customers from past purchases. We compare a collaborative filter to a recommendations engine based on geodemographic segments and the customers own history of purchases.

# Technology Operations & Information Systems: ERP Systems

Session: SB19
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: Technology Innovations & Operations
Room:
Chair: Antonio Arreola-Risa
Chair Address: Texas A&M University, Dept. of IOM,, Grad. Sch. of Bus., College Station, TX 77843-4217
Chair E-mail: tarreola@tamu.edu
Chair:
Chair E-mail:

SB19.1 ERP's Dirty Little Secret
• Mark L. Spearman; University of Alabama, Coll. of Commerce & Bus. Admin, 366 Alston Hall, Box 870226, Tuscaloosa, AL 35487-0226; mspearman@cba.ua.edu

Thirty-five years after Joseph Orlicky developed MRP, the short cuts taken remain in the latest incarnation of manufacturing control software: ERP. While it was understandable that some corners were cut using the computer technology of the 1960s, it is inexcusable today. While it appears that the only alternative are the extremely complex 'advanced planning systems,' this does not have to be the case..

SB19.2 Empirical Study on Enterprise Information System Strategies in the Japanese Pharmaceutical Industry

In the Japanese pharmaceutical industry, efforts towards business efficiency and continuous BPR are critical issues as well as the development of new efficacious medicines. ERP is very useful for BPR and implemented in many enterprises. We take a survey research using a questionnaire and explore a state-of-the-art and problems of enterprise information system in the pharmaceutical industry.

SB19.3 An Operations Research Perspective on Internet Security Issues
• Henry C. Co; California Polytechnic & State University, Coll. of Bus., Pomona, CA 91768; hco@csupomona.edu

With the recent high-profile hackers' attack on the Internet, it is unmistakable that Internet security is the biggest concern in e-commerce. Just like other security risks in business, Internet security risk can and should be managed and minimized. We examine various Internet security issues from an OR perspective. Issues explored include Internet security risks...

SB19.4 ERP Systems: Post-Implementation Experiences & Strategies
• Antonio Arreola-Risa; Texas A&M University, Dept. of IOM,, Grad. Sch. of Bus., College Station, TX 77843-4217; tarreola@tamu.edu

There are 2 kinds of companies: those that fail miserably to implement ERP systems and those that fail miserably immediately after implementing ERP systems (Michael Hammer Seminar Foundation). We present research findings regarding post-implementation experiences of ERP systems and the post-implementation strategy of a company that recently finished the implementation of an ERP system.

# Tabu Search Applied to Military Logistics Problems

Session: SB20
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: Tabu & Scatter Search
Room:
Chair: J. Wesley Barnes
Chair Address: University of Texas, Dept. of Mech. Eng., Austin, TX 78712
Chair E-mail: wbarnes@mail.utexas.edu
Chair:
Chair E-mail:

SB20.1 A Quick-Running Routing Algorithm for UAV Reconnassance Operations
• Robert Harder; AFIT, Dept. of Operational Sci.;
• Gary Kinney; AFIT, Dept. of Operational Sci.;
• James T. Moore; Air Force Institute of Technology, Dept. of Operational Sci., 2950 P St., WPAFB, OH 54533; james.moore@afit.af.mil
• Raymond R. Hill; AFIT, Dept. of Organizational Sci.;

We present a Java-based vehicle routing algorithm using tabu search as the solution engine. A user interface is developed to allow Predator unmanned aerial vehicle operators to employ this algorithm to develop UAV routes. Empirical results based on standard test problems are presented along with details of the algorithm implementation into operational practice.

SB20.2 A Group Theoretic Tabu Search Approach to the Aerial Fleet Refueling Problem

A simplified version of the aerial fleet refueling problem may be modeled as a multi-depot VRP with time window and route length constraints. Using the symmetric group on n letters, we present a GTTS approach to this problem. The GTTS approach uses the group theoretic concepts of function composition and conjugation to construct and implement move neighborhoods...

SB20.3 Using Group Theory to Characterize Metaheuristic Search Neighborhoods
• David L. Neuway; University of Texas, Grad. Program in OR, Austin, TX 78712;
• J. Wesley Barnes; University of Texas, Dept. of Mech. Eng., Austin, TX 78712; wbarnes@mail.utexas.edu

We present some findings from recent research that help the understanding of the power and limitations of some commonly used metaheuristic search neighborhoods.

SB20.4 Tabu Search Applied to the Weapon Assignment Problem
• Christopher A. Cullenbine; AFIT, Dept. of Operational Sci.;
• Mark A. Gallagher; AFIT, Dept. of Operational Sci., 2950 P St., WPAFB, OH 45433-7765;
• James T. Moore; Air Force Institute of Technology, Dept. of Operational Sci., 2950 P St., WPAFB, OH 54533; james.moore@afit.af.mil

US Strategic Command analyzes potential nuclear weapon force structures with the WAM. A typical integer program inside WAM includes over 16 goals, 500,000 decision variables and 70,000 constraints. We demonstrate that tabu search obtains excellent solutions in about 15 minutes. This approach provides better support for nuclear policy decisions.

# Session II

Session: SB21
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: SICUP
Room:
Chair: Guntram Scheithauer
Chair Address: Dresden University of Technology, Inst. for Numerical Math, Dresden, D-01062 , Germany
Chair E-mail: scheit@math.tu-dresden.de
Chair:
Chair E-mail:

SB21.1 Using Extra Dual Cuts to Accelerate Column Generation
• Jose M. Valerio De Carvalho; Universidade do Minho, Dept. Producao e Sistemas, Braga, 4710-057 , Portugal; vc@ci.uminho.pt

We derive a family of dual cuts for the one-dimensional cutting stock problem and show how inserting a polynomial number of cuts of this family in the restricted master problem helps to reduce the number of columns generated, the number of degenerate pivots and computational time.

SB21.2 Pattern Reduction in Cutting Pattern Problems
• Horacio H. Yanasse; INPE Brazilian Space Research Institute, Av. Dos Astronautas 1758, CP 515 Sao Jose dos Campos, Sao Paulo, 12201-970 , Brazil; horacio@lac.inpe.br
• Marcelo S. Limeira; INPE Brazilian Space Research Institute, Av. Dos Astronautas 1758, CP 515 Sao Jose dos Campos, Sao Paulo, 12201-970 , Brazil; marcelo@lac.inpe.br

We propose a hybrid procedure to obtain a reduced number of patterns in cutting problems. In the first stage, we generate (good) patterns that once cut, simultaneously complete the demands of items contained in them. We reduce the problem, solve it and then pattern reduction techniques, proposed previously in the literature, are

SB21.3 Solving 1-Dimensional Cutting Stock Problems Exactly with a Cutting Plane Algorithm
• Guntram Scheithauer; Dresden University of Technology, Inst. for Numerical Math, Dresden, D-01062 , Germany; scheit@math.tu-dresden.de

A cutting plane approach for solving cutting stock problems is presented. The computation and the choice of suitable cutting planes are investigated for an algorithm based on column generation. Results of extensive computational experiments are reported.

SB21.4 withdrawn - chair request of 10/2

# Modeling Health Care Operations

Session: SB22
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Murray J. Cote
Chair Address: Trinity University, San Antonio, TX 78212-7200
Chair E-mail: mjcote@trinity.edu
Chair:
Chair E-mail:

SB22.1 Development of a Simulation & Data Visualization Tool to Assist in Strategic Operations Management in Emergency Services
• Shane Henderson; University of Michigan, IOE Dept., 1205 Beal Ave., Ann Arbor, MI 48109-2117; shane.henderson@umich.edu
• Andrew J. Mason; University of Auckland, Dept. of Eng. Science, Private Bag 92019, Auckland, , New Zealand; a.mason@auckland.ac.nz

We describe a powerful decision support tool, BartSim, that may be used to assist in the problem of emergency vehicle deployment. BartSim employs simulation to perform what-if analyses, and graphical tools to analyze both actual and simulated results. BartSim has been extensively used to assist in decision making for an ambulance organization in New Zealand.

SB22.2 Synergism between Modeling & Operational Evaluation: Examples in Health Care Reengineering
• George Miller; Vector Research, Inc., PO Box 1506, Ann Arbor, MI 48106; millerg@vrinet.com

We explore the benefits of combining operational evaluation with model-based analysis to study the reengineering of healthcare delivery. We illustrate a conceptual approach to this type of analysis with several examples drawn from previous studies in the areas of telemedicine proliferation and adoption of disease management practices.

SB22.3 Medical Application of Engineering Risk Analysis & Anesthesia Patient Risk Illustration
• Elisabeth M. Pate-Cornell; Stanford University, Dept. of MS & Eng., Stanford, CA 94305-4024; mep@leland.stanford.edu

Engineering probabilistic risk analysis methods can be used in the medical domain. We present a dynamic analysis of anesthesia accidents, relate the parameters to practitioner problems (competence and alertness) and to the system's management to assess the patient risk reduction benefits of improvements such as regular re-training and increased supervision of the residents.

SB22.4 Forecasting Hourly Bed Census for Hospital Care Units

A challenging aspect of daily hospital operations is determining appropriate staffing for a given bed census. We develop an autoregressive forecasting model to predict hourly census for a care unit, examine the forecast bias and its implication on staffing and consider the cost of forecast errors in relation to staffing.

# Simulated Annealing Applications

Session: SB23
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Bruce C. Shultes
Chair Address: University of Cincinnati, Mech., Ind. & Nuclear Eng., PO Box 210116, Cincinnati, OH 45221-0116
Chair E-mail: bruce.shultes@uc.edu
Chair:
Chair E-mail:

SB23.1 A Simulated Annealing Approach for Evaluating ANSI Form Tolerances
• Liu Hong; University of Cincinnati, Mech., Ind. & Nuclear Energy, PO Box 210116, Cincinnati, OH 45221-0116; hongl@email.uc.edu
• Bruce C. Shultes; University of Cincinnati, Mech., Ind. & Nuclear Eng., PO Box 210116, Cincinnati, OH 45221-0116; bruce.shultes@uc.edu
• Sam Anand; University of Cincinnati, Mech., Ind. & Nuclear Energy, PO Box 210116, Cincinnati, OH 45221-0116; sam.anand@uc.edu

The accurate evaluation of form tolerances as defined by the ANSI Standard requires efficient methods for solving complex nonlinear problems. We present a general simulated annealing approach for evaluating form tolerances. Empirical results for several examples are presented and compared with results from existing methods found in the literature.

SB23.2 Periodic Transportation Scheduling under Uncertainty

We discuss a logistics problem where products are distributed to customers according to a regularly periodic schedule rather than through placed orders. The goal is to optimally route the delivery vehicles taking uncertainy into account. We use a sample-average version of simulated annealing to solve the problem.

SB23.3 An Efficient Sampling Approach to Probabilistic Optimization Methods
• Ki-Joo Kim; Carnegie Mellon University, Civil & Environ. Eng. Dept., 5000 Forbes Ave., Pittsburgh, PA 15213; kijoo@andrew.cmu.edu
• Urmila Diwekar; Carnegie Mellon University, Baker Hall 129, Dept. of EPP, 5000 Forbes Ave., Pittsburgh, PA 15213; ud01@andrew.cmu.edu

Efficient simulated annealing (ESA) and stochastic annealing (ESTA) are developed by changing move generation mechanism. A novel sampling technique is used for uniform move generation resulting in shorter Markov chain length. ESA and ESTA are observed to be up to 60% more efficient than the conventional ones.

SB23.4 Using Simulation to Find Feasible Schedules with Real-Time Constraints
• Bruce C. Shultes; University of Cincinnati, Mech., Ind. & Nuclear Eng., PO Box 210116, Cincinnati, OH 45221-0116; bruce.shultes@uc.edu

Hard real-time constaints arise in the specification of embedded software creating a class of complex scheduling problems. Heuristics for constructing schedules exist, but schedules are not necessarily feasible. We present a new approach that searches for feasible schedules. This approach is compared with the current state of practice through examples.

# Material Handling Systems Integration

Session: SB24
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: Material Handling
Room:
Chair: Sonia Bartolomei-Suarez
Chair Address: University Puerto Rico, Dept. of IE, Box 9043, Mayaguez, PR 00681-9043
Chair E-mail: sonia@ece.uprm.edu
Chair:
Chair E-mail:

SB24.1 Incorporating the Impact of Order Dispatching & Material Handling in the Effectiveness of Modular Manufacturing in the Apparel Industry
• Viviana I. Cesani; University Puerto Rico, Dept. of IE, Mayaguez, PR 00681-9043;
• Veronica Melendez; University Puerto Rico, Dept. of IE, Mayaguez, PR 00681-9043;

Modular manufacturing is an integral part of a comprehensive effort designed to improve productivity in the apparel industry. It has been suggested that by switching from the traditional bundle system to manufacturing modules, companies can achieve lower throughput times and significantly reduce their WIP inventory. We examine the influence that operational factors have on the performance of modular systems...

SB24.2 Pull-Based Supply Chain Allocation for Multi-Products in Multi-Echelon Distribution Systems
• Pongchai Athikomrattanakul; Iowa State University, Ames, IA 50011;
• Pius J. Egbelu; Iowa State University, Dept. of IMSE, 2019 Black Engineering, Ames, IA 50011-2164; pegbelu@iastate.edu

We propose a primal-dual decomposition method to allocate resources in a pull-based supply chain, customer demand driven, multiple echelon distribution consisting of m manufacturing centers, n warehouses and p customers. Based on an MIP model, we decompose the original problem into a master problem and subproblem to seek solution to the distribution problem.

SB24.3 Evaluation of Quadratic Assignment Problem Solution Techniques on the Integration of Assembly Planning & Material Handling Problems
• David E. Miro-Feliciano; University Puerto Rico, Dept. of IE, Mayaguez, PR 00681-9043;
• Sonia Bartolomei-Suarez; University Puerto Rico, Dept. of IE, Box 9043, Mayaguez, PR 00681-9043; sonia@ece.uprm.edu

A computer-based assembly planning system with an imbedded SA solution algorithm was developed for solving a QAP that integrates assembly problem and material handling. The effectiveness of using SA and a search algorithm based on reactive tabu-search (RTS) are compared. Two case studies are used in a comparative analysis to determine how well RTS compares with SA...

SB24.4 A Simulation Model-Based Approach for Predictive Maintenance of Conveying Equipment
• Erasmo Lopez; University of Texas, MIE Dept., El Paso, TX 79968-0521;
• Luis R. Contreras; University of Texas, MIE Dept., El Paso, TX 79968-0521; lrcontreras@miners.utep.edu
• Arunkumar Pennathur; University of Texas, MIE Dept., El Paso, TX 79968-0521; apennathur@utep.edu

We present a simulation model based on a theoretical framework for integrating equipment design variables, i.e, equipment age, wear, life, failure and reliability distributions, etc, and variables related to operating conditions of conveying equipment, i.e, temperature, load, sound, vibration, lubricant debris, etc., both of which contribute to machine breakdown...

# Nonlinear Programming, Integer Programming & Interior-Point Algorithms

Session: SB25
Date/Time: Sunday 10:15-11:45
Track:
Cluster: Nonlinear Programming
Room:
Chair: Yin Zhang
Chair Address: Rice University, Dept. of CAAM, Houston, TX 77005
Chair E-mail: zhang@caam.rice.edu
Chair:
Chair E-mail:

SB25.1 The Sphere of Convergence of Newton's Method on Two Equivalent Systems from Nonlinear Programming
• Maria Cristina Villalobos; University of Texas, Dept. of Math, El Paso, TX 79925;
• Richard A. Tapia; Rice University, Dept. of CAAM, Box 1892, Houston, TX 77005;
• Yin Zhang; Rice University, Dept. of CAAM, Houston, TX 77005; zhang@caam.rice.edu

We study local behavior of a Newton logarithmic barrier function method and a Newton primal-dual interior-point method for the nondegenerate inequality constrained optimization problem, in particular, the radius of the sphere of convergence. Our theoretical and numerical results are clearly in favor of using Newton primal-dual methods for solving the optimization problem.

SB25.2 An Algorithm for Ellipsoidal Approximation of Polytopes with an Application in Integer Programming
• Liyan Gao; Rice University, Dept. of CAAM, Houston, TX 77005;
• Yin Zhang; Rice University, Dept. of CAAM, Houston, TX 77005; zhang@caam.rice.edu

The maximum volume ellipsoidal approximation of polytopes (MVE) has applications in many areas. We propose a practical primal-dual interior point method to solve the MVE problem and compare the performance of our method with that of Khachiyan & Todd's algorithm. We apply our maximum volume ellipsoidal approximation algorithm as a subroutine in solving the IP problem based on Lenstra's algorithm...

SB25.3 An Interior-Point Gradient Method for Optimization Problems with Simple Bounds
• Amr S. El-Bakry; ExxonMobil Upstream Research Company, PO Box 2189, Houston, TX 77252-2189;

A new first-order interior-point method for minimizing a continuously differentiable function with simple bounds will be introduced. The search direction is computed using a distance function defined on the cone of nonnegative real numbers. Preliminary numerical results will be presented.

SB25.4 Approximation of the MAXCUT Problem through an Unconstrained Optimization Problem

We formulate a relaxation of the maxcut problem as an unconstrained optimization problem. Local minima of this problem can be obtained very efficiently which, surprisingly, lead to high quality approximate solutions to the maxcut problem.

# Queueing Models

Session: SB26
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster: Stochastic Models & Applications
Room:
Chair: William L. Cooper
Chair Address: University of Minnesota, Dept. of Mech. Eng., 111 Church St. SE, Minneapolis, MN 55455-0111
Chair E-mail: billcoop@me.umn.edu
Chair:
Chair E-mail:

SB26.1 On Truncation Properties of Finite-Buffer Queues & Queueing Networks
• Xiuli Chao; NJIT, Dept. of IME, University Heights, Newark, NJ 07102; chao@megahertz.njit.edu
• Masakiyo Miyazawa; Science University of Tokyo, Dept. of Info. Sciences, Noda-City Chiba, 278 , Japan;

We show that several truncation properties of queueing systems are consequences of a simple property of censored stochastic processes. We first establish a Markov chain result and then apply it to single-server batch arrival batch service queues with finite buffers and queueing networks with finite buffers and batch movements.

SB26.2 Queues with Skorohod/Loynes Representations
• William L. Cooper; University of Minnesota, Dept. of Mech. Eng., 111 Church St. SE, Minneapolis, MN 55455-0111; billcoop@me.umn.edu
• V. Schmidt; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332;
• Richard F. Serfozo; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332; rserfozo@isye.gatech.edu

We consider processes that are regulated to stay within a finite or infinite interval. Potential inputs or outputs are disregarded when they would lead to a state outside this region. We describe the relationship of such processes to a variant of the Skorohod problem and Loynes' stationary waiting time representation.

SB26.3 Stochastic Production Planning
• German Riano; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332;
• Steven T. Hackman; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205;
• Richard F. Serfozo; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332; rserfozo@isye.gatech.edu

We will address the problem of how to find proper input rates to feed a production network to meet a given set of requirements. Random processing times and congestion are taken into account in the model.

SB26.4 Reversible Markov Processes & Spatial Queueing Systems
• Richard F. Serfozo; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332; rserfozo@isye.gatech.edu
• Xiaotao Huang; Citicorp., 1 Court Sq., 30FL/03, Long Island City, NY 11120;

In a spatial service system, the state of the system is a point process that describes the numbers of customers in subsets of the space. It evolves over time as a reversible Markov process. We characterize the stationary distribution of such systems. Basics of reversibility on uncountable spaces are included.

# Panel: The Industry Job Search for PhD Students

Session: SB27
Date/Time: Sunday 10:15-11:45
Type:
Track:
Cluster:
Room:
Chair: Russell D. Meller
Chair Address: Virginia Tech., Dept. of ISE, 250 New Engineering Bldg., Blacksburg, VA 24061
Chair E-mail: rmeller@vt.edu
Chair:
Chair E-mail:

SB27.1 Panel: The Industry Job Search for PhD Students

INFORMS members from industry will provide PhD students with an outline of the industry job search, i.e., application, interview, offer process, etc. They will also provide their guidance on some of the dos and don'ts, as well make a contrast with the academic job search.

# Networks & Graphs

Session: SB28
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: James F. Campbell
Chair Address: University of Missouri, Coll. of Bus. Admin., 8001 Natural Bridge Rd., St. Louis, MO 63121
Chair E-mail: campbell@umsl.edu
Chair:
Chair E-mail:

SB28.1 Equilibrium Modeling on Capacitated Networks
• Hamdouch Younes; CRT, 4580 Queen Mary App. 402, Montreal, Quebec, H3W 1W6 , Canada; younes@crt.umontreal.ca
• Patrice Marcotte; University of Montreal, DIRO, CP 6128, Succ. Centre-Ville, Montreal, Quebec, H3C 3J7 , Canada; marcotte@diro.umontreal.ca
• Nguyen Sang; University of Montreal, CRT, CP 6128, Succ. Centre-Ville, Montreal, Quebec, H3C 3J7 , Canada; nguyens@diro.umontreal.ca

We advocate the use of the concept of strategies and their network-theoretic representation as hyperpath, for modeling equilibrium flows in networks involving arcs with rigid capacities.Both theoretical and computational salient features of this approach will be discussed and preliminary supporting computational results presented.

SB28.2 A Cutting Plane Solution Approach to the Parallel Replacement Problem under Economies of Scale

The parallel replacement problem determines optimal timing of asset disposal and procurement decisions for a group of assets over a finite horizon under the assumption of fixed and variable costs. An integer programming formulation is presented with a cutting plane solution approach and experimental results.

SB28.3 Error Bounded Heuristic Algorithms for Large-Scale Dynamic Programming
• Matthew D. Bailey; University of Michigan, 1950 Traver Rd., Apt. 206, Ann Arbor, MI 48105; mdbailey@engin.umich.edu
• Robert L. Smith; University of Michigan, Dept. of IOE, Ann Arbor, MI 48109-2117; rlsmith@umich.edu

We propose and analyze a generic heuristic algorithm that successively eliminates feasible solutions. At each step, it is guaranteed that the remaining solution set has a solution within an a priori error from the optimal. This is illustrated using dynamic programming with guaranteed error bounds.

SB28.4 Solving Hub Arc Location Problems
• James F. Campbell; University of Missouri, Coll. of Bus. Admin., 8001 Natural Bridge Rd., St. Louis, MO 63121; campbell@umsl.edu
• Andreas Ernst; CSIRO, Math. & IS, Private Bag 10, South Bank MDC, Clayton VIC, 3169 , Australia; andreas.ernst@cmis.csiro.au
• Mohan Krishnamoorthy; CSIRO, Math. & IS, Private Bag 10, South Bank MDC, Clayton VIC, 3169 , Australia; mohan.krishnamoorthy@cmis.csiro.au

Hub arc location problems involve designing hub networks by locating arcs with reduced unit flow costs. We present integer programmingformulations for several models with different path and network constraints. We also report computational results for 2 optimal solution approaches.

SB28.5 Cyclic Schedules for Kirkman Problems
• Gerald L. Thompson; Carnegie Mellon University, 5000 Forbes Ave., Pittsburgh, PA 15213-3890; gt04@andrew.cmu.edu

I will present algorithms for generating cyclic schedules to Kirkman school girl problems with n = 12k + 6 where k>0 is an integer and for other Kirkman problems with n = 6k+1. The problem of calculating the widths of the resulting matrices (in the sense of Fulkerson and Ryser) will also be discussed.

# Modeling Price Formation & Market Power in Competitive Electricity Markets I

Session: SB29
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Yves Smeers
Chair Address: Universite Catholique de Louvain, CORE, 34 Voie du Roman Pays, Louvain la Neuve, 1348 , Belgium
Chair E-mail: smeers@core.ucl.ac.be
Chair:
Chair E-mail:

SB29.1 A Focal Point Model of Tacit Collusion in Electricity Markets
• John Bower; London Business School, Sussex Place, Regent's Park, London, NW1 4SA , England, UK; jbower@london.edu
• Derek W. Bunn; London Business School, Sussex Place, Regents Park, London, NW1 4SA , England, UK; dbunn@london.edu

A common characteristic of many electricity markets is that an oligopoly of generating firms who regularly compete to supply power has ample opportunities for tacit collusion. An agent-based simulation model is presented where competing generating firms endogenously coordinate upon, and enforce, a collusive equilibrium using feedback from the capital market to select a focal point based on plant utilization rates.

SB29.2 Simulation of Price Behavior in Electricity Markets
• Ezra Hausman; Tabors Caramanis & Associates, 50 Church St., Cambridge, MA 02138;
• Aleksandr Rudkevich; Tabors Caramanis & Associates, 50 Church St., Cambridge, MA 02138; arudkevich@tca-us.com
• Assef Zobian; Tabors Caramanis & Associates, 50 Church St., Cambridge, MA 02138;

We present the results and analyses of simulating hour-to-hour behavior of locational electricity prices. Our algorithms combine a 2-stage commitment-dispatch SFE game, LP-driven secured dispatch with ramping constraints and adaptive learning based on market observation. The results are compared to observed price spikes in US electricity markets.

SB29.3 Supply Function vs. Cournot Equilibrium Models in Electricity Markets
• Ross Baldick; University of Texas, Dept. of Elect. & Comp. Eng., Austin, TX 78712-1084; baldick@ece.utexas.edu

We investigate the properties of supply function and Cournot equilibria in transmission constrained electricity markets. Several examples suggest that although supply function equilibria fit typical electricity pool rules, from a computational point of view, Cournot equilibria may be provide a better estimate of the 'focal equilibrium.'

SB29.4 Experiments with Market Power Models
• Cecile Cordier; ELECTRABEL RD Energy Markets, 34 Voie du Roman Pays, Louvain-la-Neuve, 1348 , Belgium;
• Olivier Daxhelet; ELECTRABEL RD Energy Markets, 34 Voie du Roman Pays, Louvain-la-Neuve, 1348 , Belgium;
• Yves Smeers; Universite Catholique de Louvain, CORE, 34 Voie du Roman Pays, Louvain la Neuve, 1348 , Belgium; smeers@core.ucl.ac.be

We consider different stylized situations of restructured electric systems constructed on the basis of real data. Using these models, we test the impact of different assumptions of market organization, transmission pricing and system sizes and parameters on various economic performances of the system, i.e, welfare, price levels, congestion rents, etc.

# Carl Harris Memorial Session: Optimization in Computational Biology

Session: SB30
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Harvey J. Greenberg
Chair Address: University of Colorado, Mathematics Dept., CB 170, PO Box 173364, Denver, CO 80217-3364
Chair E-mail: hgreenbe@carbon.cudenver.edu
Chair:
Chair E-mail:

SB30.1 Opportunities for Optimization in Computational Biology
• Harvey J. Greenberg; University of Colorado, Mathematics Dept., CB 170, PO Box 173364, Denver, CO 80217-3364; hgreenbe@carbon.cudenver.edu

There has been tremendous growth in this dynamic field of computational biology, creating new challenges in combinatorial and global optimization. We will describe some of these problems from an OR perspective.

SB30.2 Using Global Optimization to Predict Molecular Structure
• Richard Byrd; University of Colorado, Comp. Science Dept., Boulder, CO 80302; richard@cs.colorado.edu
• Betty Eskow; University of Colorado, Comp. Science Dept., Boulder, CO 80302;
• Bobby Schnabel; University of Colorado, Comp. Science Dept., Boulder, CO 80302;

Finding the minimum energy configuration of a protein is an important tool for understanding them and is a structured but very difficult global optimization problem. We describe a global optimization method involving random sampling, structured perturbation and a pool of candidate minimizers that is effective on such problems.

SB30.3 Combinatorial Protein Structure Problems

Although protein structures can be naturally described with continuous parameters, combinatorial descriptions of protein structures naturally arise in many contexts. We describe performance-guaranteed approximation algorithms for protein structure prediction in a variety of combinatorial models and illustrate how the analysis of these algorithms can be extended to provide off-lattice performance guarantees.

# Recent Advances in Nonlinear Mixed Integer Optimization

Session: SB31
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Arne S. Drud
Chair Address: ARKI Consulting & Development A/S, Bagsvaerdvej 246A, Bagsvaerd, DK2880 , Denmark
Chair:
Chair E-mail:

SB31.1 Recent Developments in Disjunctive 0/1 Programming
• Ignacio E. Grossmann; Carnegie Mellon University, Chemical Eng. Dept., Pittsburgh, PA 15213-3890; grossmann@cmu.edu
• Aldo Vecchietti; INGAR, Inst. de Desarrollo y Diseno, Avellanda 3657, Santa Fe, 3000 , Argentina; aldovec@arcride.edu.ar

In recent years, disjunctive programming has shown several advantages at the levels of modeling and solving discrete/continuous optimization problems. We will present recent developments on nonlinear disjunctive 0/1 models. We will also introduce several issues about language implementation for posing a problem and also the algorithms and techniques to reach the solution.

SB31.2 SBB: A New Solver for Mixed Integer Nonlinear Programming
• Michael R. Bussieck; GAMS Development Corp., 1217 Potomac St. NW, Washington, DC 20007; mbussieck@gams.com
• Arne S. Drud; ARKI Consulting & Development A/S, Bagsvaerdvej 246A, Bagsvaerd, DK2880 , Denmark; adrud@arki.dk

We describe SBB, a new GAMS solver for mixed integer nonlinear programming. The solver uses existing NLP solvers in a B&B scheme. To improve reliability, solver failures or infeasibilities are overcome by changing solver or options dynamically. Results and comparisons with other methods will be presented.

SB31.3 Global Optimization of Mixed Integer Nonlinear Programs

Towards the global optimization of MINLPs we develop: convex extensions, the only available systematic means of convexification of LSC functions; a linear outer-approximation scheme for factorable programs; a framework for domain reduction which produces existing and new range-reduction techniques. We present computational results with our BARON software.

# The Institutional Foundations of Markets

Session: SB32
Date/Time: Sunday 10:15-11:45
Track:
Cluster:
Room:
Chair: Andrew Spicer
Chair Address: , 347 Cold Sold Rd., Princeton, NJ 08540
Chair E-mail: andrew.spicer@ucr.edu
Chair:
Chair E-mail:

SB32.1 Markets as Social Institutions: Implications for Firm Strategy & Economic Development

We examine the formation of impersonal markets of exchange in the Russian banking industry during the 1990s. The challenges in forming mechanisms of exchange based on price represent a major issue for firm strategy in Russia as well as for governmental attempts to jumpstart economic growth.

SB32.2 Investments in Pricing Capital
• Mark J. Zbaracki; University of Pennsylvania, Mgmt. Dept., The Wharton Sch., 2000 SH-DH, Philadelphia, PA 19104-6370; zbaracki@wharton.upenn.edu

We argue that pricing is an output of organizational production processes, making both the costs and benefits of pricing endogenous choices. We discuss how firms choose pricing rules and heuristics, thereby linking organizational processes to economic variables. We support our argument with evidence from a 2-year, cross-disciplinary, ethnographic study.

SB32.3 The Institutional Construction of the US Mutual Fund
• Michael Lounsbury; Cornell University, NYSSILR, 367 Ives Hall, Ithaca, NY 14853-3901; mdl18@cornell.edu

As sociologists have increasingly turned their attention to the study of economic institutions over the past couple of decades, we have gained a great deal of knowledge about how markets are comprised of a complex infrastructure of rules, relationships and resources that enable stable patterns of exchange. Much less attention has been paid to market dynamics...

SB32.4 Unpacking Affiliation Markets: Partner Search & Selection in Airline
• M. Tina Dacin; Texas A&M University, Dept. of Mgmt., Grad. Sch. of Bus., College Station, TX 77843; tdacin@tamu.edu

I study the selection and competition for exchange partners in airline alliance networks. Drawing upon the literature on networks and economic sociology, I show how attributes such as reputation, status, prior ties and alliance histories serve as important mechanisms of stratification and contribute to the creation of markets of affiliation.

# Production Planning I

Session: SB33
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Raf Jans
Chair Address: Katholieke Universiteit Leuven, Dept. of Applied Econ., Naamsestraat 69, Leuven, 3000 , Belgium
Chair E-mail: raf.jans@econ.kuleuven.ac.be
Chair:
Chair E-mail:

SB33.1 An Approximate Dynamic Programming Technique for Production
• Meko M. C. So; University of Hong Kong, Dept. of Math., Pokfulam Rd., Hong Kong, , Hong Kong; h9622462@hkusua.hku.hk
• Michael K. P. Ng; University of Hong Kong, Dept. of Math., Pokfulam Rd., Hong Kong, , Hong Kong; mng@maths.hku.hk

We develop a dynamic programming model for the production planning. An approximate solution is determined by searching for a lower bound via LP and reducing the complexity of the search space by a priority products scheme. Numerical results are reported on a set of illustrative data.

SB33.2 Job Sequencing in CONWIP Production Lines
• Wen Zhang; Concordia University, Dept. of Mech. Eng., 1455 de Maisonneuve West, Montreal, Quebec, H3G 1M8 , Canada; zhang_1@me.concordia.ca
• Mingyuan M. Chen; Concordia University, Dept. of Mech. Eng., 1455 de Maisonneuve West, Montreal, Quebec, H3G 1M8 , Canada; mychen@me.concordia.ca

We discuss some issues in using CONWIP control policies to manufacturing production lines. A mathematical programming model is proposed to simultaneously determine the most concerned essential parameters in CONWIP controlled manufacturing system. Some related issues such as bottleneck production smoothness and line production smoothness are also discussed.

SB33.3 Service Level Relationships in Multi-Stage Production Systems under Base Stock Policy
• Subrata Mitra; Indian Institute of Management, Fellow Programme Office, PO 16757, Alipore PO, Calcutta West Bengal, 700 027 , India; subratamitra@hotmail.com
• Ashis K. Chatterjee; Indian Institute of Management, PO 16757, Alipore PO, Calcutta West Bengal, 700 027 , India; ac@iimcal.ac.in

In integrated production-distribution systems, the customer service level is a function of service levels at different stages. We address the problem of determining the stage-specific service levels to achieve a desired customer service level in multi-stage production systems under base stock policy. Both optimal and heuristic procedures have been considered.

SB33.4 A Production Planning Problem in the Tire Industry

We present an industrial production planning problem with start up times at a tire manufacturer. We solve problems with up to 30 products and 30 periods by a column generation scheme. The master is solved using Lagrange relaxation. We further discuss the impact of our planning program.

# Decision Analysis II

Session: SB34
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Robert F. Nau
Chair Address: Duke University, Fuqua Sch. of Bus., Durham, NC 27514
Chair E-mail: robert.nau@duke.edu,, http://www.duke.edu/~rnau
Chair:
Chair E-mail:

SB34.1 withdrawn - author request of 10/9

SB34.2 Roulette Revisited: Martingales, Risk Attitudes & Rationality

By considering risk preferences, I rationalize the demonstrated irrationality of playing the well known martingale strategy in the game of roulette. It characterizes necessary and sufficient conditions under which a utility maximizing decision maker would play the martingale. These conditions suggest that many such utility maximizing gamblers may choose to play the strategy...

SB34.3 no show
• Xia Pan; University of Rhode Island, Coll. of Bus., Ballentine 344A, Kingston, RI 02881; xpan0114@postoffice.uri.edu
• David Shao; University of Rhode Island, Dept. of MIE, Kingston, RI 02881;

SB34.4 A Generalization of Pratt-Arrow Measures to Non-Expected-Utility Preferences

A generalized measure of local risk aversion is developed for choices under uncertainty with non-expected-utility preferences, inseparable subjective probabilities and utilities and unobserved stochastic prior wealth. As illustrations, the Ellsberg & Allais paradoxes are explained using a model of 'partitionable' utility that exhibits local uncertainty aversion everywhere, unlike Choquet expected utility.

# Forecasting II

Session: SB35
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Kwang-Soo Lee
Chair Address: Indiana State University, Analytic Dept., Terre Haute, IN 47809
Chair E-mail: sdkwang@befac.indstate.edu
Chair:
Chair E-mail:

SB35.1 withdrawn - chair request of 9/18

SB35.2 Forecasting for Dual Technology Diffusion

Many areas of technological development are characterized by distinct generations of technology, where the new generation is more efficient than its predecessor. We consider the case where the new generation is the form of a merged product of dual functions of existing technology. Examples include VOD which combines the function of cable TV and Internet service...

SB35.3 Bayesian Analysis of a Discrete Time Poisson Model with Applications to Call Volume Data
• Refik Soyer; George Washington University, Monroe Hall 403, MS Dept., 2115 G St. NW, Washington, DC 20052; soyer@gwu.edu
• Murat Tarimcilar; George Washington University, 2115 G St. NW, Washington, DC 20052; soyer@gwu.edu

We consider a discrete time series model based on Poisson counts where the intensity function includes seasonal as well as covariate effects. Our model can be considered as a discrete version of the modulated Poisson models of Cox (1972). We develop Bayesian analysis for the model using Markov chain Monte Carlo methods and apply it to some real life call volume data.

SB35.4 Moving Regression

In parallel with the moving average, a moving regression model is tested. This model generates the centered moving average as an interpolation and provides multi-period forecasts as extrapolations. Using the quarterly consumer credit outstanding data, accuracy of the model fitting and forecast is investigated as compared to the moving average.

# Inventory Management II

Session: SB36
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Margaret K. Schaefer
Chair Address: College of William & Mary, 125 Jones Hall, Dept. of Math, Williamsburg, VA 23187-7895
Chair E-mail: mkscha@math.wm.edu
Chair:
Chair E-mail:

SB36.1 An Inventory Model with Two Vendors & Random Yield

We consider a single item inventory system with 2 vendors in single period. Yields are random, due to random capacity at vendor. A model is proposed to find the optimal order quantity for each vendor that will minimize the total expected costs of purchasing, inventory holding and shortage.

SB36.2 Exact Evaluation of Service Performance for a Periodic Review, Two-Echelon Inventory System

We consider analyzing service performance for a stochastic, periodic review, 2-echelon inventory system. The system consists of a supplier and a group of downstream retailers who independently replenish their inventory according to respective periodic review, lot-size reorder point policies. The supplier orders from an exogenous source and supplies the retailers which in turn meets customers...

SB36.3 Order Coordination in a Two-Echelon Distribution System with Random Demand at Retailers
• Joong Y. Son; University of Washington, MS Dept., Box 353200, Seattle, WA 98195; sonjy@u.washington.edu
• Ted D. Klastorin; University of Washington, Dept. of MS, Box 353200, Seattle, WA 98195-3200; tedk@u.washington.edu
• Kamran Moinzadeh; University of Washington, Dept. of MS, Sch. of Bus., Box 353200, Seattle, WA 98195; kamran@u.washington.edu

We analyze an order coordination scheme in a 2-echelon decentralized supply chain with random demands at retailers where coordination is achieved through a pricing and order timing mechanism. We propose an extended (R,T) policy for the retailers and demonstrate the effectiveness of the coordination policies throughout the supply chain.

SB36.4 The Multi-Location Inventory Centralization Problem with First-Come, First-Served Allocation
• Amit Eynan; Washington University, 1 Brookings Dr., CB 1133, St. Louis, MO 63130; eynan@olin.wustl.edu

Each location may be characterized by a unique selling price and unique consequences of stockouts, reflecting its market economic structure. Although inventory allocation (and revenue) is contingent upon customers arrival processes generated in various locations, a single location equivalent can be used. The impact on profit and service level is presented.

SB36.5 Video Inventory Systems: The Blockbuster Scenario
• Margaret K. Schaefer; College of William & Mary, 125 Jones Hall, Dept. of Math, Williamsburg, VA 23187-7895; mkscha@math.wm.edu
• Keith E. Laba; Fidelity Investments, 109 Hillside Rd., #2, Watertown, MA 02472; keith.laba@fmr.com

Recently, Blockbuster Videos has entered into agreements with major movie studios, under which Blockbuster purchases videos at deeply discounted prices in return for rebating part of their rental revenues. Thus Blockbuster purchases many copies of new releases and reduces stockout probabilities. We modify our video inventory model for the Blockbuster scenario.

# Optimization Techniques II

Session: SB37
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Renyou Wang
Chair Address: Carnegie Mellon University, Porter Hall, Dept. of EPP, 5000 Forbes Ave., Pittsburgh, PA 15213
Chair E-mail: renyou@andrew.cmu.edu
Chair:
Chair E-mail:

SB37.1 An Integrated Approach for Railway Crew Planning
• Paolo Toth; University of Bologna, DEIS, Viale Risorgimento 2, Bologna, 40136 , Italy; ptoth@deis.unibo.it
• Alberto Caprara; University of Bologna, DEIS, Viale Risorgimento 2, Bologna, 40136 , Italy; acacaprara@deis.unibo.it
• Michele Monaci; University of Bologna, DEIS, Viale Risorgimento 2, Bologna, 40136 , Italy; mmonaci@deis.unibo.it

Given a set of daily train services, find a min-cost set of crew rosters covering them. Generally the problem is solved through 3 phases: pairing generation, pairing optimization and rostering optimization. We describe an effective technique integrating pairing and rostering optimizations. Computational results on real-world Italian Railways instances are presented.

SB37.2 Optimization with Neural Networks Trained by Genetic Algorithms
• Marta I. Velazco Fontova; UNICAMP, Sch. Elect. Eng./Comp. Sci., Av. A. Einstein 400, CxP 6101, Campinas SP, 13083-970 , Brazil; velazco@densis.fee.unicamp.br
• Christiano Lyra; UNICAMP, Sch. Elect. Eng. & Comp. Sci., Av. A. Einstein 400, CxP 6101, Campinas SP, 13083-970 , Brazil; chrlyra@densis.fee.unicamp.br

Multi-layer neural networks are trained to approach optimization problems. GAs are adopted to 'evolve' weights, unveiling new points in the definition domain. As evolution goes on, better points are found, driving the process to optimality. Case studies compare the new method with other neural network approaches to optimization.

SB37.3 On the Convergence of the Weiszfeld Algorithm
• Halit Uster; Texas A&M University, Dept. of Info. & Op. Mgmt., MS 4217, College Station, TX 77843-4217; uster@tamu.edu
• Robert F. Love; McMaster University, Sch. of Bus., 1280 Main St. West, Hamilton, Ontario, L8S 4M4 , Canada; rlove@mcmaster.ca

We investigate the convergence properties of the Weiszfeld procedure when it is applied to the approximated Lp-norm single-facility location problem where p>2. We show that convergence for p>2 can be obtained by introducing a step size factor to the iterative procedure. Some numerical test results are also given.

SB37.4 Latin Hypercube Hammersley Sequence Sampling: A New & Efficient Sampling Technique for Optimization under Uncertainty
• Renyou Wang; Carnegie Mellon University, Porter Hall, Dept. of EPP, 5000 Forbes Ave., Pittsburgh, PA 15213; renyou@andrew.cmu.edu
• Urmila Diwekar; Carnegie Mellon University, Baker Hall 129, Dept. of EPP, 5000 Forbes Ave., Pittsburgh, PA 15213; ud01@andrew.cmu.edu

We will present a new and efficient sampling technique, the Latin Hypercube-Hammersley sequence sampling technique, by coupling Latin hypercube sampling and Hammersley sequence sampling techniques. The distribution-free approach to inducing correlation of uncertainties will also be presented. This sampling technique will be a great help for optimization under uncertainties.

# NLP II

Session: SB38
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Michael M. Kostreva
Chair Address: Clemson University, Dept. of Math. Sci., Clemson, SC 29634
Chair E-mail: flstgla@clemson.edu,, http://www.math.clemson.edu/faculty/Kostreva/mfd.html
Chair:
Chair E-mail:

SB38.1 Nonlinear Optimization without Derivatives

We are interested in the problem of minimizing an objective function of several variables whose derivatives are unavailable. Instances of this problem occur in engineering, and typically the number of variables is moderate and the objective is very expensive to compute. We present a numerical algorithm for solving this problem.

SB38.2 withdrawn - author request of 10/27
• Mehmet Tolga Cezik; Columbia University, Mudd Bldg., Rm. 318, 500 West 120th St., New York, NY 10027; cezik@ieor.columbia.edu
• Garud Iyengar; Columbia University, Mudd Bldg., Rm. 314, 500 West 120th St., New York, NY 10027; garud@ieor.columbia.edu

SB38.3 Exact Solution of the Subproblem in the Cutting Angle Method for Optimization of Abstract Convex Functions

The cutting angle method for optimization of some classes of abstract convex functions is a version of the bundle method or can be considered as an extension of the cutting plane method of convex optimization. It is an iterative method. At each iteration, a nonconvex optimization subproblem is solved. Computational complexity of the method is essentially defined by the complexity of this subproblem...

SB38.4 A Probabilistic Method of Feasible Directions

A probabilistic version of the MFD for solving NLP problems with inequality constraints is presented. Randomization is applied to modify the multiple direction MFD algorithm. Some numerical experiments on problems with known solutions serve to compare this method with the traditional deterministic versions of MFD.

# Marketing II

Session: SB39
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Imran S. Currim
Chair Address: University of California, Grad. Sch. of Mgmt., Irvine, CA 92697-3125
Chair E-mail: iscurrim@uci.edu
Chair:
Chair E-mail:

SB39.1 withdrawn - author request of 10/16

SB39.2 A Study for an Allocation of Sales Promotion Costs
• Tabata Tomoaki; Waseda University, Sch. of Science & Eng., 3-4-1 Okubo, Shinjuku, Tokyo, 169-8555 , Japan; tabata@mgmt.waseda.ac.jp
• Namatame Takashi; Science University of Tokyo, Dept. of IME, Fac. of Eng., 1-3 Kagurazaka, Shinjuku, Tokyo, 162-8601 , Japan; namatame@ms.kagu.sut.ac.jp
• Takahiro Ohno; Waseda University, Ohno Lab. 3-4-1, Okubo Shinjuku, Tokyo, 169-8555 , Japan; ohno@ohno.mgmt.waseda.ac.jp

It is an important subject for a retail store to get and maintain profitable prime customers. Therefore, promotions for customers need to be differentiated in terms of contribution rate to retail store. We propose the optimal promotion allocations that maximize the long-term profit of retail store by analyzing ID-POS data.

SB39.3 Noisy Equilibrium & Neologism-Proofness in Experimental Signaling Games

The vexing problem of multiplicity of equilibria in cheap-talk games has spurred research on refinements and equilibrium selection remains an as yet unresolved theoretical issue. We test neologism-proofness, a refinement of perfect Bayesian equilibrium in experimental sender-receiver games with costless messages.

SB39.4 no show
• Imran S. Currim; University of California, Grad. Sch. of Mgmt., Irvine, CA 92697-3125; iscurrim@uci.edu
• Rick L. Andrews; University of Delaware, Dept. of Bus. Admin., Newark, DE 19716; andrewsr@udel.edu

# Reliability

Session: SB40
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Jorge Valenzuela
Chair Address: Auburn University, 211Dunstan Hall, Auburn, AL 36849
Chair E-mail: jvalenz@eng.auburn.edu,, http://www.eng.auburn.edu/~jvalenz
Chair:
Chair E-mail:

SB40.1 A Tabu Search Application to the Redundancy Allocation Problem
• Sadan Kulturel-Konak; Auburn University, Dept. of ISE, 207 Dunstan Hall, Auburn, AL 36849; sadan@eng.auburn.edu
• Alice E. Smith; Auburn University, Dept. of ISE, 207 Dunstan Hall, Auburn, AL 36849; aesmith@eng.auburn.edu
• David W. Coit; Rutgers University, Dept. of IE, 96 Frelinghuysen Rd., Piscataway, NJ 08854-8018; coit@rci.rutgers.edu

A tabu search method is introduced to effectively solve the redundancy allocation problem. An adaptive penalty approach that exploits the short term memory structure of the tabu list along with the long term memory of the search results is used and is shown to be effective on a well-known test problem from the literature.

SB40.2 Threshold Graphs & the Reliability Polynomial
• Manoj K. Chari; Louisiana State University, Dept. of Math., Baton Rouge, LA 70803; chari@math.lsu.edu

We discuss a new graph theoretic interpretation of the coefficients of the connectedness reliability polynomial of undirected graphs. We apply this to the class of threshold graphs this gives some partial results on a conjecture of Boesch et.al on the existence of uniformly least reliable graphs.

SB40.3 Maximum Reliability at Minimum Cost
• Richard M. Soland; George Washington University, Dept. of EMSE, School of Eng., Washington, DC 20052; soland@seas.gwu.edu

We treat the problem of making choices for the various subsystems of a given system in order to achieve high system reliability at low total cost. Preliminary results are presented for the case of a system consisting of independently failing subsystems in series.

SB40.4 The Effect of Power Generation Reliability on the Spot Price of Electricity

The electricity market is undergoing deregulation. Short- and long-term economic decisions will be based on the spot price of electricity. We present results of a Monte Carlo simulation used to study the effect of generating unit availabilities on the price at which electricity is traded.

# Forecasting Accounting Numbers

Session: SB41
Date/Time: Sunday 10:15-11:45
Sponsor: Accounting, Auditing & Tax Section
Track:
Cluster:
Room:
Chair: Shiva K. Sivaramakrishnan
Chair Address: Texas A&M University, Dept. of Acct., Grad. Sch. of Bus., College Station, TX 77843-4353
Chair E-mail: shiva@cgsb.tamu.edu
Chair:
Chair E-mail:

SB41.1 Are Selling, General & Administrative Costs Sticky?
• Mark C. Anderson; University of Texas, Sch. of Mgmt., PO Box 830688, MS J043, Richardson, TX 75083-0688; andersmc@utdallas.edu
• Rajiv Banker; University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO43, Richardson, TX 75083-0688; rbanker@utdallas.edu
• Surya Janakiraman; University of Texas at Dallas, Sch. of Mgmt., JO4.3, Richardson, TX 75083; suryaj@utdallas.edu

Costs are commonly assumed to vary symmetrically with changes in activity. We develop and test a hypothesis that costs are sticky - increase more with increases in activity than they decrease with decreases in activity. We document that SG&A costs exhibit sticky cost behavior.

SB41.2 Service Capacity Decision & Incentive Compatible Cost Allocation for Reporting Usage Forecasts
• Suresh Radhakrishnan; University of Texas at Dallas, Sch. of Mgmt., MS JO43, Richardson, TX 75083-0688; sradhakr@utdallas.edu
• Kashi R. Balachandran; NYU, Stern Sch. of Bus., New York, NY 10012; kbalacha@stern.nyu.edu

We consider an M/G/1 queue with multiple users. Service capacity can be improved at a cost by reducing the mean and variance of service time. The expected usage is private information of each user. We develop a simple cost allocation scheme that induces truthful reports of expected usage forecasts, maximizes the expected net benefit of each user...

SB41.3 Valuation Implications of Meeting or Beating Analysts' Earnings Forecasts
• Gia Chevis; Texas A&M University, Dept. of Acct., College Station, TX 77843-4353; gchevis@cgsb.tamu.edu
• Somnath Das; University of Illinois, Dept. of Acct., Chicago, IL; sdas@uic.edu
• Shiva K. Sivaramakrishnan; Texas A&M University, Dept. of Acct., Grad. Sch. of Bus., College Station, TX 77843-4353; shiva@cgsb.tamu.edu

Casual evidence suggests that many firms seem to be attaching importance to meeting or beating analysts' earnings expectations consistently. We use Ohlson's framework to examine whether the capital market values such firms differently. Results indicate that valuations of these firms are significantly higher relative to firms not adopting this strategy even after controlling for size, financial risk, etc...

# Dantzig Dissertation Award Finalist Presentations

Session: SB42
Date/Time: Sunday 10:15-11:45
Type: Invited
Track:
Cluster:
Room:
Chair: Zelda B. Zabinsky
Chair Address: University of Washington, Dept. of IE, Box 352650, Seattle, WA 98195-2650
Chair E-mail: zelda@u.washington.edu
Chair:
Chair E-mail:

SB42.1 Strategic Planning under Uncertainty: Stochastic Integer Programming Approaches
• Shabbir Ahmed; University of Illinois, Dept. of MIE, 1206 West Green St., Urbana, IL 61801; s-ahmed1@uiuc.edu

No abstract supplied.

SB42.2 A Threshhold Inventory Rationing Policy for Service Differentiated Demand Classes

No abstract supplied.

SB42.3 Operational & Incentive Issues in Health Care Systems

No abstract supplied.

SB42.4 Channel Policies & Supply Chain Coordination

No abstract supplied.

SB42.5 Solution of Large-Scale Allocation Problems with Partially Observable Outcomes

No abstract supplied.

# Risk Management II

Session: SB43
Date/Time: Sunday 10:15-11:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Lyn C. Thomas
Chair Address: University of Southampton, Mgmt. School, Highfield, Southampton, Hants, SO17 1BJ , UK
Chair E-mail: l.thomas@ed.ac.uk
Chair:
Chair E-mail:

SB43.1 withdrawn - author request of 9/15
• Yesim Tokat; University of California at Santa Barbara, 795 Juniper Walk #D, Goleta, CA 93117; yesim@econ.ucsb.edu
• Svetozar T. Rachev; University of California, Dept. of Stats & Applied Prob., Santa Barbara, CA 93106-3110;
• Eduardo Schwartz; UCLA, Anderson Grad. Sch. of Mgmt., 110 Westwood Plaza, Los Angeles, CA 90095-1481; eduardo.schwartz@anderson.ucla.edu

SB43.2 A Comparison of Credit Scoring Model Building Techniques

We analyze the effects of continuous vs. discrete variables and the use of several statistical techniques for credit scoring models. A comparison of these models predictive power is also presented.

SB43.3 Resetting of Segregated Fund Maturity Guarantees

We model the reset decision for the maturity guarantee of a segregated fund using a discrete time Markov chain. We characterize the optimal value for the minimum return threshold and show that the simple strategy of resetting whenever the fund's value has increased is sub-optimal but can perform well.

SB43.4 PHab Scoring: Survival Analysis for Personal Loans
• Lyn C. Thomas; University of Southampton, Mgmt. School, Highfield, Southampton, Hants, SO17 1BJ , UK; l.thomas@ed.ac.uk
• Maria Stepanova; University of Southampton, Mgmt. School, Highfield, Southampton, Hants, SO17 1BJ , UK;

Credit scoring has proved very successful in managing the default risk in consumer lending by assessing who is likely to default. To move to profit scoring, one needs to estimate when someone will default. Survival analysis is an ideal way of modeling this. We discuss how proportional hazards models are modified to develop proportional hazards applications to behavioral scores and proportional hazard applications to credit scores.

# COIN Installfest for Linux: A Hands-On Workshop using Open Source Software for OR

Session: SB46
Date/Time: Sunday 10:15-11:45
Type:
Track:
Cluster:
Room:
Chair Address: IBM Research, TJ Watson Research Ctr., PO Box 218, Rte. 134, Yorktown Heights, NY 10598
Chair:
Chair E-mail:

SB46.1 Installation & Use of Common Optimization INterface (COIN): Linux
• Laszlo Ladanyi; IBM Research, TJ Watson Research Ctr., PO Box 218, Rte. 134, Yorktown Heights, NY 10598; ladanyi@us.ibm.com
• J. P. Fasano; IBM Research, TJ Watson Research Ctr., Rte. 134, Yorktown Heights, NY 10598; jpfasano@us.ibm.com
• Robin Lougee-Heimer; IBM TJ Watson Research Center, PO Box 218, Rte. 134, Yorktown Heights, NY 10598-0218; robinlh@us.ibm.com
• Matthew J. Saltzman; Clemson University, Dept. of Math. Sci., Clemson, SC 29634-0975; mjs@math.clemson.edu

COIN is an initiative to promote open source software for OR. Bring your Linux laptop and network card to participate in this hands-on workshop using open source components in the COIN repository. No Linux-laptop? Come observe and learn more about COIN. Visit http://oss.software.ibm.com/developerworks/opensource/coin for details and prerequisites.

# Medical Decision Making

Session: SC01
Date/Time: Sunday 13:15-14:45
Track:
Cluster:
Room:
Chair: Scott B. Cantor
Chair Address: University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Box 40, Houston, TX 77030-4095
Chair E-mail: sbcantor@mdanderson.org
Chair:
Chair E-mail:

SC01.1 Impact of Patient & Provider Preferences on the Cost-Effectiveness of Therapy for Recurrent Rectal Carcinomas
• Alexander R. Miller; University of Texas Health Science Center, Dept. of Surgery, 7703 Floyd Curl Dr., San Antonio, TX 78289; millerar@uthscsa.edu
• Scott B. Cantor; University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Box 40, Houston, TX 77030-4095; sbcantor@mdanderson.org
• George E. Peoples; University of Texas, Anderson Cancer Ctr., 1515 Holcombe Blvd., Box 40, Houston, TX 77030-4095;
• David B. Pearlstone; University of Texas, Anderson Cancer Ctr., 1515 Holcombe Blvd., Box 40, Houston, TX 77030-4095;
• John M. Skibber; University of Texas, Anderson Cancer Ctr., 1515 Holcombe Blvd., Box 40, Houston, TX 77030-4095;

We performed a cost-effectiveness analysis of therapeutic options for patients with locally recurrent rectal carcinoma. We created a decision-analytic model that incorporated outcomes of survival, quality of life and costs. Utilities were elicited from convenience samples of 24 health care providers and 24 patients using the standard gamble technique.

SC01.2 Time-Tradeoff Utility Assessment with Equal Horizons: Would You Prefer 'Good then Bad' or 'Moderate'?
• Robert M. Hamm; University of Oklahoma Health Sciences Center, Family & Preventive Med. Dept., 900 NE 10th St., Oklahoma City, OK 73104; robert-hamm@ouhsc.edu

I propose a variation of the time trade-off method for utility assessment. Patients are asked to choose between a moderately unhappy life or a life of the same length that is first happy, then unhappy. Adjustments for temporal discounting can be made. Data from 50 patients will be presented.

SC01.3 Decision Making for the Treatment of Early Stage Prostate Cancer
• Brian J. Miles; Baylor College of Medicine, Scott Dept. of Urology, 1 Baylor Plaza, Houston, TX 77030; bmiles@bcm.tmc.edu

Predictive modeling is commonly used in business decisions, occasionally in clinical research and rarely in the clinic. The use of modeling is highly sensitive to probabilities, utilities and assumptions, often supporting contradictory positions. Nonetheless, decision analysis for treatment of early stage prostate cancer could be promising for selective patients.

SC01.4 Prostate Cancer Screening Recommendations based on Couple's Utilities
• Scott B. Cantor; University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Box 40, Houston, TX 77030-4095; sbcantor@mdanderson.org
• Robert J. Volk; Baylor College of Medicine, Family & Community Med. Dept., Scurlock Tower, Ste. 1406, Houston, TX 77030; bvolk@bcm.tmc.edu
• Murray D. Krahn; Toronto Hospital, Toronto, Ontario, , Canada; murray.krahn@utoronto.ca
• Alvah R. Cass; University of Texas Medical Branch, Dept. of Family Med., 301 University Blvd., Galveston, TX 77555-1123; acass@utmb.edu
• Stephen J. Spann; Baylor College of Medicine, Family & Community Med. Dept., 5510 Greenbriar, Rm. 266, Houston, TX 77005; sspann@bcm.tmc.edu

We incorporated patient ages and husbands', wive's and couples' utilities into a decision-analytic model of prostate cancer screening. Subjects were 168 couples from 3 family pracice centers. Screening was recommended for 48 husbands, 89 wives and 58 couples. The recommendations showed greatest concordance for men's and couple's utilities.

Session: SC02
Date/Time: Sunday 13:15-14:45
Track:
Cluster:
Room:
Chair: Dana R. Clyman
Chair Address: University of Virginia, Darden Grad Sch. of Bus. Adm., PO Box 6550, Charlottesville, VA 22906-6550
Chair E-mail: clymand@darden.virginia.edu
Chair:
Chair E-mail:

SC02.1 Recent Decision Analysis Applications in Operations Research Literature
• Donald L. Keefer; Arizona State University, Dept. of Supply Chain Mgmt., Tempe, AZ 85287-4706; don.keefer@asu.edu
• Craig W. Kirkwood; Arizona State University, Dept. of Supply Chain Mgmt., Tempe, AZ 85287-4706; craig.kirkwood@asu.edu
• James L. Corner; University of Waikato, Dept. of Mgmt. Systems, Private Bag 3105, Hamilton, , New Zealand; jcorner@waikato.ac.nz

We survey applications of decision analysis that appeared in major English language OR journals and other closely related journals from 1990 through 1999. This is an update to the earlier survey of decision analysis applications from 1970 through 1989 by Corner & Kirkwood.

SC02.2 What's Cooking at the Lab?: Resource Allocation Decisions in Basic Science
• David Reiter; Stanford University, Dept. of MS/Eng., Sch. of Eng., Terman Eng. Ctr. 324, Stanford, CA 94305-4023;

The federal government will spend over 20 billion on basic science research in 2001. These resources are typically allocated via informal processes that allow emotion, politics and personal biases to play large roles. A decision analytic approach adds structure, insight and clarity to the process. SC02.3 A Measure of Relevance • Roberto Ley-Borras; Instituto Tecnologico de Orizaba, Oriente 13 A No. 1122, Entre Norte 28 y 28 A, Orizaba, Veracruz, 94380 , Mexico; ley@stanfordalumni.org We introduce a single real number measure of relevance (probabilistic dependence) between discrete uncertain variables. The measure is easy to compute, is consistent with our intuitive assessment of relevance, can enhance the representation of relevance in influence diagrams and can improve communication between participants in the decision process. SC02.4 (Most) Disagreements among Experts are Illusory What seems to be a disagreement among experts about the state of the world is often just a result of vague language, reference to different questions or partial information. A more interesting source of apparent disagreement is a common methodological failure with regard to the role of experts in the evaluation of uncertain situations. # Supply Contracts & Information Session: SC03 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: MSOM Track: Cluster: Room: Chair: Alexander O. Brown Chair Address: Vanderbilt University, Owen Grad. Sch. of Mgmt., 401 21st Ave. South, Nashville, TN 37203 Chair E-mail: alex.brown@owen.vanderbilt.edu Chair: Chair Address: Chair E-mail: SC03.1 Channel Conflict & Supply Chain Coordination • Andy A. Tsay; Santa Clara University, Leavey Sch. of Bus., Dept. of OMIS, Santa Clara, CA 95053-0382; atsay@scu.edu • Narendra Agrawal; Santa Clara University, Leavey Sch. of Bus., Dept. of OMIS, Santa Clara, CA 95053-0382; Recent developments in Internet-based commerce have led many companies to consider direct sales. Such a company may at once be both a supplier to and a direct competitor of existing reseller partners, resulting in 'channel conflict'. We model key attributes of this setting, and examine ways to improve channel relationships. SC03.2 A Bayesian Inventory Problem with Price-Dependent Demand & Lost Sales • Martin A. Lariviere; Duke University, Fuqua Sch. of Business, Box 90120, Durham, NC 27705; larivier@mail.duke.edu Consider a setting with stochastic, price-dependent demand and lost sales. Some parameter of the distribution is unknown, and beliefs about the parameter are updated in a Bayesian fashion. Pricing and stocking decisions consequently affect both profits and learning. We examine how decisions are adjusted to gather information. SC03.3 Theory & Practice of Shared Savings Contracts in Supply Chains • Gregory A. DeCroix; Duke University, Fuqua Sch. of Bus., Box 90120, Durham, NC 27705; • Charles J. Corbett; UCLA, AGSM, 110 Westwood Plaza, Box 951481, Los Angeles, CA 90095-1481; charles.corbett@anderson.ucla.edu In a 'shared savings' contract, supplier and customer are both given incentives to reduce consumption of the material involved. Such shared savings contracts are becoming increasingly common in some sectors, notably (but not exclusively) in chemical management services for the automotive industry. In this paper we provide an overview of current state of the art of the practice of such contracts and their effects, and show how these effects can be modeled. SC03.4 Pricing Supply Flexibility We consider a model of a manufacturer and a supplier of a long lead time, expensive product. The supplier currently offers its product at a single price with 0 supply flexibility. We consider how the supplier should price additional supply flexibility when the manufacturer has private information about demand. # Inventory Policies with Demand Curves Session: SC04 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: MSOM Track: Cluster: Room: Chair: Apostolos N. Burnetas Chair Address: Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106-7235 Chair E-mail: atb4@po.cwru.edu Chair: Chair Address: Chair E-mail: SC04.1 Newsvendor Policies for Demand-Stimulating Inventories • Euthemia Stavrulaki; Pennsylvania State University, 303 Beam Bldg., MSIS Dept., Smeal Coll. of Bus., University Park, PA 16802-1913; exs37@psu.edu • Anantaram Balakrishnan; Pennsylvania State University, 303 Beam Bldg., MSIS Dept., Smeal Coll. of Bus., University Park, PA 16802-1913; anantb@psu.edu • Michael Pangburn; Pennsylvania State University, 315 Beam Bldg., MSIS Dept., Smeal Coll. of Bus., University Park, PA 16802-1913; mikepangburn@psu.edu,, msp5@psu.edu For certain products (e.g., fashion goods), larger inventories stimulate greater retail demand. We model the influence of stocking levels on demand distribution, and show that the profit-maximizing quantity, obtained using a modified critical fractile relationship, exceeds the equilibrium quantity ordered by a newsvendor who does not explicitly recognize demand stimulation. SC04.2 Global Production Planning under Exchange-Rate Uncertainty & Price Sensitivity • Burak Kazaz; Loyola University, IS & Op. Mgmt. Dept., 25 East Pearson, Ste. 1400, Chicago, IL 60611; bkazaz@luc.edu We study a multinational company producing both at home and abroad. The firm dynamically adjusts the prices for its products as the exchange rates fluctuate - this alters the demand for its products as well. We provide all the potentially optimal production policies and the conditions that lead to them. SC04.3 Pricing Policies to Influence the Ordering Decisions of Downstream Retailers • Viswanath Cvsa; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106; vxc10@po.cwru.edu We analyze the flexibility in pricing for a monopolist manufacturer who approaches the market through a network of competing retailers. We identify policies that permit the manufacturer to influence the timing of the retailers' orders. We also examine the consequences of these policies for the performance of the supply chain. SC04.4 Option Contracts in Supply Chains • Peter Ritchken; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106; phr@po.cwru.edu • Apostolos N. Burnetas; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106-7235; atb4@po.cwru.edu We investigate the pricing of option contracts when the demand curve is downward sloping. A manufacturer offers the retailer reorder and/or return options. We examine the consequences of these contracts on the wholesale price and the volatility of the retail price and their impact on both agents' profits. # Analysis & Control of Queueing Systems Session: SC05 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Applied Probability Society Track: Cluster: Room: Chair: Hayriye Ayhan Chair Address: Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205 Chair E-mail: hayhan@isye.gatech.edu Chair: Chair Address: Chair E-mail: SC05.1 A Diffusion Approximation for Queues with Reneging • Amy L. Ward; Stanford University, Dept. of MS & Eng., Terman Engineering Ctr., Stanford, CA 94305-4026; award@leland.stanford.edu • Peter W. Glynn; Stanford University, Dept. of MS & Eng., Terman Engineering Ctr., Stanford, CA 94305-4026; glynn@leland.stanford.edu We develop a diffusion approximation for a queueing system that includes job deadlines. We propose a nonstandard approximation: an Ornstein-Uhlenbeck process with additional downwards drift and reflection at the origin. We further provide explicit formulas for estimating important system performance measures and compare these estimations with exact results. SC05.2 Increasing Flexibility • Mark E. Lewis; University of Michigan, IOE Dept., 1205 Beal Ave., Ann Arbor, MI 48109-2117; melewis@engin.umich.edu Consider a controlled M/M/1/K system where customers may be subject to 2 assessments. The first occurs upon arrival and the second, which occurs with some probability p, occurs when the customer reaches the front of the queue. The initial decision-maker has limited knowledge of the customer's work requirements. SC05.3 Exact Asymptotics for Rare Events in a Queueing System with Multiple Customer Types • Jerome D. Coombs-Reyes; Georgia Institute of Technology, School of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205; jerome@isye.gatech.edu • Robert D. Foley; Georgia Institute of Technology, School of ISyE, Atlanta, GA 30332-0205; rfoley@isye.gatech.edu Consider 2 queues and 4 customer types. Two are dedicated, the third joins the shorter queue and the fourth, the queue with the shorter expected waiting time. We obtain exact asymptotic expressions for the stationary probability of and the expected time until some large level of customers. SC05.4 Laplace Transform & Moments of Waiting Times in Poisson Driven (max, +)-Linear Systems • Hayriye Ayhan; Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205; hayhan@isye.gatech.edu • Dong-Won Seo; Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205; dongwon@isye.gatech.edu (Max,+) linear systems can be used to model a class of queueing networks. We provide explicit expressions for the moments and the Laplace transform of transient waiting times in Poisson driven (max,+) linear systems. Furthermore, starting with these closed form expressions, we also derive explicit expressions for the moments and the Laplace transform of stationary waiting times... # Enhancing Travel Forecasting Models Session: SC06 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Transportation Science Section Track: Cluster: Room: Chair: Hani S. Mahmassani Chair Address: University of Texas, Dept. of Civil Eng., ECJ 6.2, Austin, TX 78712-1076 Chair E-mail: masmah@mail.utexas.edu Chair: Chair Address: Chair E-mail: SC06.1 A Planning Model for Urban Intermodal Transportation Networks • Khaled F. Abdelghany; University of Texas, Dept. of Civil Eng., ECJ 6.2, Austin, TX 78712; kfaissal@mail.utexas.edu • Hani S. Mahmassani; University of Texas, Dept. of Civil Eng., ECJ 6.2, Austin, TX 78712-1076; masmah@mail.utexas.edu We present a planning and operation decision tool for urban intermodal transportation networks. The modeling framework includes multi-objective dynamic traffic assignment with discrete stochastic route-mode choice capabilities. Various modeling issues together with a set of experiments that illustrate the functionality of the model are presented. SC06.2 Household Level Commute Travel Mode Choice & Allocation of Activity Stops Effects of allocation of activity stops within a household on travel mode choice for the work commute is investigated using the 1996 Bay Area Travel Survey. Past research on commute stops has focused on individual mode choice, not on joint decisions of travel mode choice and activity stop-making at the household level. SC06.3 A Simplicial Decomposition Method for the Dynamic Network Equilibrium Problem • Jose M. Rubio-Ardanaz; University of Montreal, CRT, CP 6128, Succ. Centre-ville, Montreal, Quebec, H3C 3J7 , Canada; jose@crt.umontreal.ca • Jia Hao Wu; INRO Consultants, Inc., 5160 Decarie Blvd., Ste. 610, Montreal, Quebec, H3X 2H9 , Canada; wujiahao@crt.umontreal.ca • Michael A. Florian; University of Montreal, CRT, CP 6128, Succ. Centre-Ville, Montreal, Quebec, H3C 3J7 , Canada; mike@crt.umontreal.ca In our iterative method for solving the analytical dynamic network equilibrium problem, at each iteration, we solve the dynamic network equilibrium problem over a simplicial space and then find the time-dependent shortest route in order to increase the simplicial space. The FIFO condition is needed and we show that it is satisfied in our model. Numerical results are given. SC06.4 Enhancements of the Gradient Projection Algorithm for the Traffic Assignment Problem • Der-Horng Lee; Naational University of Singapore, Dept. of Civil Eng., Singapore, 117576 , Singapore; dhl@nus.edu.sg • Yu Nie; Naational University of Singapore, Dept. of Civil Eng., Singapore, 117576 , Singapore; • Anthony Chen; Utah State University, Dept. of Civil & Environ. Eng., Logan, UT 84322-4110; We computationally show the weakness of the GP algorithm with regard to its accuracy and stability. Possible improving strategies such as step size optimization, application of conjugate gradient and cyclic flow update are considered to enhance the performance of GP. Numerical results are reported. # Helping Handicapped Students Overcome the Unique Challenges they Face in OR/MS Courses Session: SC07 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: INFORM-ED Track: Cluster: Room: Chair: James J. Cochran Chair Address: Louisiana Tech. University, Dept. of Comp. IS & Anlysis, Coll. of Admin. & Bus., Ruston, LA 71272 Chair E-mail: cochran@cab.latech.edu Chair: Chair Address: Chair E-mail: SC07.1 Academic Support Services for the Disabled at Wright State University Extending educational opportunities to people with disabilities has been a high priority at Wright State University since its 1967 founding. We discuss the mission of Academic Support Services, especially the Technology Center, which provides classroom materials in alternative formats that include audio cassette tapes, computer disks, braille and image enhancement. SC07.2 Teaching OR/MS to Visually Impaired Students • Ana I. Aviles; Northwestern University, Dept. of IE/MS, 2145 Sheridan Rd., MEAS C210, Evanston, IL 60208; ivelisse@iens.nwu.edu The challenges that professors face when teaching a technical subject to a blind or low vision student are incredible. This presentation involves methods of teaching OR/MS to visually impaired students, the attitudes about the potential of people with disabilities, some legal aspects and funding opportunities. # Financial Management II Session: SC08 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Ngome Ntoko Chair Address: SUNY, Sch. of Bus., Oswego, NY 13126 Chair E-mail: ntoko@oswego.edu Chair: Chair Address: Chair E-mail: SC08.1 Do Stocks Perform Better in the Second Half of a US Presidential Administration? • Donald N. Stengel; California State University, 5245 North Backer Ave., MS PB07, Fresno, CA 93740-8001; don_stengel@csufresno.edu An article in the AAII journal observed that the average annual return for the S&P 500 in the third and fourth year of an administration exceeded the average return in the first 2 years, based on 1946-1998 data. We apply standard hypothesis tests to these data, with surprising results. SC08.2 Pricing Treasury Inflation Protected Securities & Related Derivatives using an HJM Model We use an HJM model to price TIPS and related derivative securities. This is a 4-step process. First, using the market prices of TIPS and ordinary US Treasury securities, both the TIPS and nominal 0-coupon bond price curves are obtained using standard coupon-bond price stripping procedures. Real 0-coupon bond price curves are then computed from the TIPS 0-coupon bond prices... SC08.3 Selection of Stocks using Value Criteria: Can Neural Network Models Yield Superior Risk Adjusted Returns? • Stanley R. Stansell; East Carolina University, Sch. of Business, Greenville, NC 27858; stansells@mail.ecu.edu • Stanley Eakins; East Carolina University, Sch. of Business, Greenville, NC 27858; eakinss@mail.ecu.edu We seek to determine whether superior risk adjusted investment returns can be obtained using neural network forecasting models. We cover a 20-year period. The results indicate that, using commonly accepted measures of value, the selected stock portfolios yield risk adjusted returns superior to the DJIA, the SP500 and a group of randomly selected stocks. SC08.4 no show • Mehrdad Tamiz; University of Portsmouth, Sch. of Comp. Sci. & Math., Mercantile House, Portsmouth, PO1 2EG , UK; mehrdad.tamiz@port.ac.uk SC08.5 A Dynamic Pricing Model of Optimal Pricing & Investment with Market & Finance Constraints We investigate dynamic pricing by a firm that needs to price competitively to secure market share as well as high enough to generate profits required for future cost-reducing investment. A model is developed that reconciles these seemingly contradictory strategies while maximizing the long-run dividends to shareholders. # Pricing & Revenue Management Session: SC09 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Revenue Management Section/Aviation Applications Section Track: Cluster: Room: Chair: Georgia Perakis Chair Address: MIT, Sloan Sch. of Bus., OR Ctr., 50 Memorial Dr., Rm. E53-359, Cambridge, MA 02139 Chair E-mail: georgiap@mit.edu Chair: Chair Address: Chair E-mail: SC09.1 Single-Leg Yield Management under a Discrete Choice Model of Demand • Garrett J. van Ryzin; Columbia University, 412 Uris Hall, Grad. School of Bus., New York, NY 10027; gjv1@columbia.edu • Kalyan Talluri; Universitat Pompeu Fabra, 27 Ramon Trias Fargas, Barcelona, 08005 , Spain; kalyan.talluri@econ.upf.es We analyze a single leg yield management problem in which customer demand is modeled using discrete choice theory. We show a nested allocation structure is optimal. We also propose an estimation procedure based on the EM method. Numerical results are presented. SC09.2 A Bi-Level Programming Model of Revenue Mangaement in the Airline Industry • Patrice Marcotte; University of Montreal, DIRO, CP 6128, Succ. Centre-Ville, Montreal, Quebec, H3C 3J7 , Canada; marcotte@diro.umontreal.ca • Gilles Savard; Ecole Polytechnique, Dept. of Math. & IE, Box 6079 Centre-Ville, Montreal, Quebec, H3C 3A7 , Canada; gilles@crt.umontreal.ca We propose a bi-level model for determining jointly optimal pricing and seat allotment policies for an airline. The main feature of our approach is that customers optimize their own 3-criterion objective over a set of alternatives provided by both the optimizing company and its competitors. Numerical approaches are presented. SC09.3 Simulation-Based Optimization for Airline Network Revenue Management • Sanne de Boer; MIT, OR Ctr., E40-130, 1 Amherst St., Cambridge, MA 02139; sanne@mit.edu • Dimitris Bertsimas; MIT, Sloan Sch. of Mgmt., E53-359, OR Ctr., 50 Memorial Dr., Cambridge, MA 02139; dbertsim@mit.edu The predominant methods to airline network RM utilize deterministic mathematical programming formulations. We combine approximate dynamic programming ideas and simulation based optimization to explicitly address the stochastic and dynamic character of demand. In computational experiments this approach leads to significant and practically feasible improvements over current practice. SC09.4 A Dynamic Pricing Approach in Revenue Management • Georgia Perakis; MIT, Sloan Sch. of Bus., OR Ctr., 50 Memorial Dr., Rm. E53-359, Cambridge, MA 02139; georgiap@mit.edu • Dimitris Bertsimas; MIT, Sloan Sch. of Mgmt., E53-359, OR Ctr., 50 Memorial Dr., Cambridge, MA 02139; dbertsim@mit.edu We take an optimization approach to pricing in a dynamic environment. We present results in a noncompetitive as well as competitive environment. Our approach does not assume the demand as known but rather as part of the model. Our results apply in a number of application areas including the airline, service and retail industries. # Optimization & Scheduling I Session: SC10 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Scheduling Room: Chair: Lei Lei Chair Address: Rutgers University, Dept. of MS/IS, Sch. of Mgmt., 180 University Ave., Newark, NJ 07102 Chair E-mail: llei@andromeda.rutgers.edu Chair: Chair Address: Chair E-mail: SC10.1 Set-Up Coordination between Two Stages of a Supply Chain • Alessandro Agnetis; Universita di Siena, Dipt. di Ingegneria & Info., via Roma 56, Siena, 53100 , Italy; agnetis@dii.unisi.it,, http://www.dii.unisi.it/~agnetis • Paolo Detti; Universita di Roma Tre, Dipt. Informatica & Auto, Roma, , Italy; • Carlo Meloni; Universita di Roma Tre, Dipt. Informatica & Auto, Roma, , Italy; • Dario Pacciarelli; Universita di Roma Tre, Dipart. di Info. & Automazione, via della Vasca Navale 84, Roma, 00146 , Italy; Part batches, of different shapes and colors, flow through a plant. In the first (second) department, a set-up occurs every time the shape (color) of the batch changes. The problem is to sequence the batches to minimize the total number of set-ups. An efficient heuristic approach is presented. SC10.2 On the Lampshade Packing Problem • Thomas Davoine; Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854-8003; davoine@rutcor.rutgers.edu • Gabriela Alexe; Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854; alexe@rutcor.rutgers.edu • Sorin Alexe; Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854; salexe@rutcor.rutgers.edu • Endre Boros; Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854-8003; boros@mozart.rutgers.edu • Peter L. Hammer; Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854-8003; hammer@rutcor.rutgers.edu • Chawalit Jeenanunta; Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854-8003; • Lei Lei; Rutgers University, Dept. of MS/IS, Sch. of Mgmt., 180 University Ave., Newark, NJ 07102; llei@andromeda.rutgers.edu • Albert Williams; Rutgers University, RUTCOR, 640 Bartholomew Rd., New Brunswick, NJ 08901; We present a problem of optimal packing coming from industrial context and requiring the partitioning of sets of objects of unusual geometric shape into a minimum number of subsets, each of which can be packed into a single box. A heuristic method is presented along with computational results. SC10.3 Shorten the Gaps to Practical Scheduling: A Survey on Reactive Scheduling • Nelson Liu; HKUST, Dept. of IEEM, Clear Water Bay, Kowloon, , Hong Kong; ieypliu@ust.hk • Benjamin Yen; HKUST, Dept. of IEEM, Clear Water Bay, Kowloon, , Hong Kong; benyen@ust.hk Along the evolution of reactive scheduling study, it has been recognized that there exists a gap between application demand and theoretical research. Some scholars adopt an application-oriented tool for problem solving, while others use more profound modeling and analysis methods to study the properties of scheduling stability and robustness. We will present the existing reactive scheduling systems for industrial applications... SC10.4 Assessing Web Site Impact on Companies: A DEA Approach • Shanling Li; McGill University, Faculty of Mgmt., 1001 Sherbrooke St. West, Montreal, Quebec, H3A 1G5 , Canada; li@management.mcgill.ca • Zhili Ouyang; Chinese Academy of Science, Inst. of MS, Beijing, , China; We use DEA to study the efficiency of e-commerce by studying numerous web sites of companies. Given that most websites are characterized by qualitative features, we develop an approach to synthesize the media richness in websites. We will discuss methodologies and computational results. # Boolean Methods in Discrete Optimization Session: SC11 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Integer Programming Room: Chair: Endre Boros Chair Address: Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854-8003 Chair E-mail: boros@mozart.rutgers.edu Chair: Peter L. Hammer Chair Address: Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854-8003 Chair E-mail: hammer@rutcor.rutgers.edu SC11.1 A Linear Time Algorithm for the Recognition of a Modular Set • Jan C. Bioch; Erasmus University Rotterdam, Dept. of Comp. Sci., FEW H9-30, Rotterdam, 3000 DR , The Netherlands; bioch@few.eur.nl Modular sets have been studied by researchers in reliability theory, game theory (committees) and Boolean function theory (Ashenhurst decomposition). Ramamurthy has shown that recognition of a modular set of a positive Boolean function with n variables and m prime implicants can be done in time O(n m^2) and that the modular closure of a set can be found in time O(n^2 m^2)... SC11.2 Finding all Minimal Feasible Solutions • Endre Boros; Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854-8003; boros@mozart.rutgers.edu • Vladimir Gurvich; ; • Leo Khachiyan; ; • Kaz Makino; ; We show that the number of maximal infeasible integer solutions to a monotone system of inequalities is linearly bounded by the number of minimal feasible solutions, implying that checking the completeness of a set of minimal feasible solutions is equivalent with dualization, and hence it is unlikely to be NP-hard. SC11.3 Dynamic Branch & Bound & Resolution Search • Fred W. Glover; University of Mississippi, Hearin Ctr. for Enterprise Sci, Sch. of Bus. Admin., University, MS 38677; fglover@bus.olemiss.edu • Said Hanafi; Universite de Valenciennes et du Hainaut Cambresis, LAMIH-UMR CNRS No. 8530, Le Mont Houy, BP 311, Valenciennes Cedex, 59304 , France; Resolution search, an innovative approach to pure 0-1 IP problems recently proposed by Chvatal, provides a novel alternative to implicit enumeration. Surprisingly, an overlooked general MIP method, dynamic B&B, yields the same branching strategies as resolution search and other strategies in addition. We suggest refinements to both. SC11.4 Maximal Bi-Cliques & Pseudo-Boolean Optimization After showing how maximal (not necessarily induced) complete bipartite subgraphs of a graph can be used for simplifying pseudo-Boolean optimization problems, we will show how all these subgraphs can be generated in total polynomial time and conclude with results of computational experiments illustrating the simplifying power of the proposed method. # Research & Industry Reviews: SCM Best Practices Session: SC12 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Logistics & Supply Chain Management Room: Chair: Yasemin Aksoy Chair Address: Tulane University, Cnsrtm. for Supply Chain Mgmt., Freeman Sch. of Bus., New Orleans, LA 70118 Chair E-mail: yasemin.aksoy@tulane.edu Chair: Chair Address: Chair E-mail: SC12.1 International Taxonomic Review of Supply Chain Management Research: A German & Scandanavian Answer to Ganeshan et al. (1999) • Herbert Kotzab; Copenhagen Business School, Dept. of Op. Mgmt., Solbjerg Plads 3, Frederiksberg, DK-2000 , Denmark; hk.om@cbs.dk The paper refers to the research review of supply chain management presented by Ganeshan et al. (1999). We broaden their view using a German and Scandinavian research perspective. Integrating these schools of SCM-thinking to the general discussions should help to identify a global valid understanding of SCM. SC12.2 withdrawn - chair request of 11/1 • Richard Dawe; Fritz Institute of Global Logistics, 4415 Cowell Rd., Ste. 200C, Concord, CA 94518-1945; rick.dawe@fritz.com SC12.3 Transformation of Third Party Logistics Services • Anu H. Bask; Helsinki School of Economics & Business Administration, Dept. of Logistics, Runeberginkatu 22-24, Helsinki, 00100 , Finland; abask@hkkk.fi Emerging e-commerce and networked markets call for new structures of outsourcing and coordination of logistics services. We discuss the transforming roles of third party logistics providers and services they offer to support different buyer-seller relationships, in particular, the need for logistics integrator. SC12.4 How OR Contributes to the Third Party Logistics Industry • Yasemin Aksoy; Tulane University, Cnsrtm. for Supply Chain Mgmt., Freeman Sch. of Bus., New Orleans, LA 70118; yasemin.aksoy@tulane.edu We review the strategic role of the third party logistics (3PL) industry in globalization of firms and discuss how operations research contributes to the 3PL industry. # Implementations of Supply Chain Design Session: SC13 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Manufacturing & Logistics Room: Chair: Marc Goetschalckx Chair Address: Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205 Chair E-mail: Chair: Chair Address: Chair E-mail: SC13.1 The Logistics of Offshore Manufacturing • Russell D. Meller; Virginia Tech., Dept. of ISE, 250 New Engineering Bldg., Blacksburg, VA 24061; rmeller@vt.edu • Pamela H. Vance; Emory University, Goizueta Bus. Sch., 1300 Clifton Rd. NE, Atlanta, GA 30322; pamela_vance@bus.emory.edu Offshore manufacturing can change the logistics system within your company. A stochastic program is presented that indicates how the manufacturing and logistics system should be configured for an offshore manufacturing option. The impact on the logistics system is presented for a company that moved part of its its manufacturing operations offshore. SC13.2 Production-Distribution System Design: Insights from an Implementation • Vedat Verter; McGill University, Faculty of Mgmt., 1001 Sherbrooke St. West, Montreal, Quebec, H3A 1G5 , Canada; verter@management.mcgill.ca • Abdullah Dasci; McGill University, Fac. of Mgmt., 1001 Sherbrooke St. West, Montreal, Quebec, H3A 1G5 , Canada; adasci@management.mcgill.ca We will present a model for simultaneous optimization of the plant location, sizing and technology decisions. The focus will be on the insights we obtained via the implementation of our formulation on a realistic production-distribution system design problem. SC13.3 Designing Supply Chain Systems for International Manufacturing Operations under Fixed Markup • Marc Goetschalckx; Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205; • Carlos J. Vidal; University del Valle, Dept. Prod. e Investigacion, Apartado Aereo 25840, Cali, , Colombia; carlos77@isye.gatech.edu We will present a model capable of designing an optimal global supply chain for manufacturing operations, when the local country operations sell goods to each other on a cost plus a fixed profit margin principle. We will discuss iterative solution algorithm and report our computational experience from an industrial implementation. # EPD Mini-Cluster: Web/IT Driven Product Development II Session: SC14 Date/Time: Sunday 13:15-14:45 Type: Invite/Sponsor Sponsor: Technology Management Section Track: Cluster: Product Development Room: Chair: Nitin Joglekar Chair Address: Boston University, Sch. of Mgmt., 595 Commonwealth Ave., Boston, MA 02215 Chair E-mail: joglekar@bu.edu Chair: Chair Address: Chair E-mail: SC14.1 Benefitting from Task-Redefining Process Technology: The Influence of an Organizational Process on the Introduction of 3D-CAD • Yaichi Aoshima; Hitotsubashi University, Inst. of Innovation Research; • Kentaro Nobeoka; Research Institute for Economics & Business Administration, 2-1 Rokkodaicho Nada-ku Kobe, Hyogo, 657-0013 , Japan; nobeoka@rieb.kobe-u.ac.jp • Yoko Takeda; International University of Japan, Ctr. for Global Communication, Harks Roppongi Bldg. 2F, Tokyo, 106-0032 , Japan; yoko@glocom.ac.jp By using questionnaire data obtained from 169 Japanese machinery-related manufacturing companies, we examined a hypothesis of the means by which organizations adapt to, and benefit from, the 3D-CAD technology which we characterized as 'task-redefining process technology' depends largely upon the organizational processes by which it is introduced. SC14.2 How does Information Technology Impact Product Development Performance? • Vish V. Krishnan; University of Texas, Dept. of Mgmt., CBA 4.202, Austin, TX 78712; krishnan@mail.utexas.edu • Viswanthan Veerkar; University of Texas, Austin, TX 78712; We consider the impact of IT on development productivity. We present data from a major computer manufacturer that shows how a web-based software helps the company manage risk and cycle time. The potential of the technology to personalize development approach and integrate with suppliers is also discussed. SC14.3 PIM Enablers for Product Development Collaboration across the Extended Enterprise • Manu Vedapudi; The Alpha Building, 15303 South Commerce Park, Ste. 100, Dearborn, MI 48120; The rapidly changing business environment, particularly in the automotive industry, is causing a shift in business models between partner organizations from one where the emphasis is on data exchange, to one where it will be on product data collaboration. We propose PIM enablers to support this paradigm shift. # Panel: Supply Chain Management - Everything Including the Kitchen Sink Session: SC15 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Quantitative Models in Supply Chain Management Room: Chair: Michael Magazine Chair Address: University of Cincinnati, QAOM Dept., PO Box 210130, Cincinnati, OH 45221-0130 Chair E-mail: mike.magazine@uc.edu Chair: Chair Address: Chair E-mail: SC15.1 Panel: Supply Chain Management - Everything, Including the Kitchen Sink • Nikhil Jain; University of Cincinnati, Coll. of Bus. Admin., PO Box 210230, Cincinnati, OH 45221; nikhil.jain@uc.edu • Steven Nahmias; Santa Clara University, Dept. of OMIS, Leavey Sch. of Bus., Santa Clara, CA 95053; • Jayashankar Swaminathan; University of California, Berkeley; msj@haas.berkeley.edu • Sridhar Tayur; Carnegie Mellon University, Grad. Sch. of Indust. Admin., Pittsburgh, PA 15213; stayur@cyrus.andrew.cmu.edu,, stayur@andrew.cmu.edu • Paul Zipkin; Duke University, Fuqua Sch. of Bus., Durham, NC 27708-0120; Supply chain management is no longer simply multi-echelon inventory, logistics and location models. It includes outsourcing, coping with product variety, e-commerce and many things that are as much pop culture as SCM. One wonders if this will have a detrimental effect on this field and it will become the TQM of the 00s. # Weighting & Ranking Decision Methods Session: SC16 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: MCDM Room: Chair: Stelios H. Zanakis Chair Address: Florida International University, Coll. of Business Admin., Dept. of DS/IS, Miami, FL 33199 Chair E-mail: Chair: Chair Address: Chair E-mail: SC16.1 An Analysis of Similarity of Some Consensus Ranking Methods • Stelios H. Zanakis; Florida International University, Coll. of Business Admin., Dept. of DS/IS, Miami, FL 33199; • Solomon Antony; Texas Tech. University, Coll. of Business Admin., ISQS Area, Lubbock, TX 79409; santony@ttu.edu We examine the behavior of 12 methods for aggregating an ordinal decision matrix. The rows may be alternative proposals or candidates, and the columns may represent different voters or criteria. This type of problem arises in several MCDA methods, in voting situations or when constraints force an ordinal scale in rating (ranking) alternatives. SC16.2 New Attribute-Weighting Techniques for Multi-Attribute Decision Making • F. Fred Choobineh; University of Nebraska, Dept. of IE, Lincoln, NE 68588-0518; fchoobineh@unl.edu • Patrick J. McKenna; US Air Force, US Strategic Command, Offutt AFB, NE; mckennap@statcom.mil In a typical MADM situation, decision makers identify feasible alternatives, select the relevant attributes, determine attributes values and assign weights to attributes. We present 4 new attribute-weighting techniques that are based on rough set theory and compare them with existing techniques. Numerical examples will demonstrate techniques use. SC16.3 Selecting an E-Commerce Strategy using Multi-Attribute Utility Analysis • Gary L. Stading; Texas A&M University, ETID Dept., College Station, TX 77843-3367; stading@entc.tamu.edu • Daniel F. Jennings; Texas A&M University, ETID Dept., College Station, TX 77843-3367; jennings@entc.tamu.edu • F. Barry Lawrence; Texas A&M University, ETID Dept., College Station, TX 77843-3367; lawrence@entc.tamu.edu • Robert J. Vokurka; Texas A&M University, ETID Dept., College Station, TX 77843-3367; vokurka@entc.tamu.edu • Gail M. Zank; Texas A&M University, ETID Dept., College Station, TX 77843-3367; zank@entc.tamu.edu The processes by which senior managers formulate strategies developing a competitive advantage are not well understood. We investigate the process which a diversified major electrical-electronics firm formulated an e-commerce strategy to attain a competitive advantage. MAUT was chosen as the method to arrive at an encompassing view of developing e-commerce strategies. # Network Flows: Applications in Transportation Session: SC17 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Optimization Section Track: Cluster: Network Flows Room: Chair: Ravindra K. Ahuja Chair Address: University of Florida, Dept. of ISE, 303 Weil Hall, Gainesville, FL 32611 Chair E-mail: ahuja@ise.ufl.edu,, ahuja@ufl.edu Chair: Chair Address: Chair E-mail: SC17.1 Solving Weighted-Constrained Network Design Problems • Anantaram Balakrishnan; Pennsylvania State University, 303 Beam Bldg., MSIS Dept., Smeal Coll. of Bus., University Park, PA 16802-1913; anantb@psu.edu • Esra Ozaygen; Pennsylvania State University, 303 Beam Bldg., University Park, PA 16802-1913; We investigate modeling and solution approaches to the weight-constrained network design problem. This problem, a generalization of hop-constrained network design, seeks the minimum cost network configuration and multi-commodity routes subject to upper limits on the total weight or length of the path used for each commodity. The model has applications in location, transportation and telecommunications. SC17.2 Optimization of Goods Flows in Supply Chains • H. Edwin Romeijn; University of Florida, Dept. of ISE, 303 Weil Hall, PO Box 116595, Gainesville, FL 32611-6595; romeijn@ise.ufl.edu • Dolores Romero Morales; Erasmus University, Rotterdam, , The Netherlands; • Richard Freling; Erasmus University Rotterdam, Econometric Inst., PO Box 1738, Rotterdam, 3000 DR , The Netherlands; • Albert P. M. Wagelmans; Erasmus University Rotterdam, Econometrics Inst., PO Box 1738, Rotterdam, DR 3000 , The Netherlands; We consider the problem of minimizing total costs in a supply chain network consisting of plants, warehouses and customers. In an environment where demand is not constant over time, we obtain a dynamic model where both transportation and inventory decisions are integrated. Heuristic and exact approaches are compared. SC17.3 Solving the Airline Combined Through & Fleet Assignment Problem using Very Large-Scale Neighborhood Search • James B. Orlin; MIT, Sloan Sch. of Mgmt., OR Ctr., Cambridge, MA 02139; jorlin@mit.edu • Ravindra K. Ahuja; University of Florida, Dept. of ISE, 303 Weil Hall, Gainesville, FL 32611; ahuja@ise.ufl.edu,, ahuja@ufl.edu • Dushyant Sharma; MIT, OR Ctr., Cambridge, MA 02139; In airline scheduling, the fleet assignment model is an integer program in which the objective is to optimally assigning a fleet of planes to flight legs. The ctFAM is the fleet assignment problem that also takes into account increased revenues from through flights. We solve ctFAM using very large-scale neighborhood search techniques. SC17.4 Solving Large-Scale Locomotive Scheduling Problems • Ravindra K. Ahuja; University of Florida, Dept. of ISE, 303 Weil Hall, Gainesville, FL 32611; ahuja@ise.ufl.edu,, ahuja@ufl.edu • Anurag Chandra; MIT, OR Ctr., Cambridge, MA 02139; • Ozlem Ergun; MIT, OR Ctr., 77 Mass Ave., Bldg. E40-130, Cambridge, MA 02139; ozie@mit.edu • Liu Jian; University of Florida, Dept. of ISE, 303 Weil Hall, Gainesville, FL 32611; • Niklas Johansson; KTH, Stockholm, , Sweden; • Carl D. Martland; MIT, Dept. of Civil & Environ. Eng., Rm. 1-153, Cambridge, MA 02139; • James B. Orlin; MIT, Sloan Sch. of Mgmt., OR Ctr., Cambridge, MA 02139; jorlin@mit.edu • Dushyant Sharma; MIT, OR Ctr., Cambridge, MA 02139; This presentation stems from an OR application at CSX Transportation, one of the major railroads in the US. The problem considered consists of assigning locomotives to trains so as to minimize the total cost of assignment while satisfying a variety of practical constraints and business rules. We will discuss our progress-to-date in solving this problem. # Issues in Internet Marketing Session: SC18 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: E-Commerce Room: Chair: P. K. Kannan Chair Address: University of Maryland, Smith School of Bus., Dept. of Mktg., College Park, MD 20742-1815 Chair E-mail: pkannan@rhsmith.umd.edu Chair: Chair Address: Chair E-mail: SC18.1 The Impact of the Internet on the Search for Automobiles • Brian T. Ratchford; University of Maryland, Smith Sch. of Bus., College Park, MD 20742; • Myung-Soo Lee; SUNY, Buffalo, NY; • Debebrata Talukdar; SUNY, Buffalo, NY; We study the impact of the Internet on consumers' for automobiles based on survey data from consumers in the Northeast. We compare the results with our previous findings that focused on consumers' general search behavior to highlight the impact of the Internet. SC18.2 no show • S. Balachander; University of Maryland, Smith Sch. of Bus., College Park, MD 20742; SC18.3 Pricing of Internet Security Products: The Implication of Network Effects • P. K. Kannan; University of Maryland, Smith School of Bus., Dept. of Mktg., College Park, MD 20742-1815; pkannan@rhsmith.umd.edu • Judy Frels; University of Maryland, Smith Sch. of Bus., Van Munching Hall, College Park, MD 20742; jfrels@rhsmith.umd.edu We present a model for pricing software-based products to provide security for Internet-based transactions. We present a discussion of the factors that need to be considered in developing such a model including network effects which increases the effectiveness of the product in providing security and privacy. # Electronic Business: Issues & Trends Session: SC19 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Technology Innovations & Operations Room: Chair: Henry C. Co Chair Address: California Polytechnic & State University, Coll. of Bus., Pomona, CA 91768 Chair E-mail: hco@csupomona.edu Chair: Chair Address: Chair E-mail: SC19.1 The New E-Business Rules are the Old Direct Marketing Rules • Robert Schaffer; ; The recent tumult on the NASDAQ and other exchanges has focused attention on the reasons why some e-businesses are successful and the reasons why so many others seem to fail. We look at some basic direct marketing theories and apply that theory to e-business. In particular, we examine 3 concepts: customer acquisition, customer retention and customer life-time value... SC19.2 E-Business & Enterprise Resource Planning: Their Relationship in the Marketplace & in the Curriculum • Lynn Turner; ; Enterprise-wide information systems such as SAP, PeopleSoft and Oracle have had to make major adjustments to meet the expectations of the movement toward e-business. We review current trends in e-business and the role that ERP systems are likely to play in this new business model. We assess the relationship of academic initiatives in enterprise computing and in e-business. SC19.3 The Use of Enterprise System Software in Undergraduate Curriculums: An Empirical Study • Randy Guthrie; ; In the past 2 years, the 3 major ES software vendors, SAP, PeopleSoft and Oracle have initiated special university alliance programs which are intended to affordably bring ES software learning into the business and/or IT curriculum of undergraduate universities. More than 100 universities are currently participating or are planning to adopt one of the 3 packages in the near future... SC19.4 Current Research Topics in Business-to-Business • Mei Qi; ; Orbitz, Covisint, Transora are just a few examples of the new fast emerging e-marketplaces for business purchasing transactions among buyers and suppliers. Buyers and suppliers are jumping in to initiate, adopt, joint venture this new emerging capable B2B technology. The emerging e-marketplaces promises to increase competition through cost reduction; yet they also brought the attention of the FTC and the DOJ... # Metaheuristics for Vehicle Routing Session: SC20 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Tabu & Scatter Search Room: Chair: Michel Gendreau Chair Address: DIRO/Universite de Montreal, CRT, CP 6128, succ. Centre-ville, Montreal, Quebec, H3C 3J7 , Canada Chair E-mail: michelg@crt.umontreal.ca Chair: Chair Address: Chair E-mail: SC20.1 A Unified Tabu Search Heuristic for Vehicle Routing Problems with Time Windows • Jean-Francois Cordeau; University of Montreal, Ecole des HEC, 3000 Chemin Cote-Ste-Catherine, Montreal, Quebec, H3C 3J7 , Canada; • Gilbert Laporte; University of Montreal, Ecole des HEC, 3000 Chemin Cote-Ste-Catherine, Montreal, Quebec, H3C 3J7 , Canada; gilbert@crt.umontreal.ca • Anne Mercier; University of Montreal, Ecole des HEC, 3000 Chemin Cote-Ste-Catherine, Montreal, Quebec, H3C 3J7 , Canada; We present a unified tabu search heuristic for the VRPTW, the periodic VRPTW and the multi-depot VRPTW. The method is quick, simple and flexible. Computational results on benchmark and random instances confirm the quality of the approach. SC20.2 Cooperative Parallel Strategies for the Vehicle Routing Problem with Time Windows • Alexandre le Bouthillier; University of Montreal, DIRO & CRT, CP 6128, Succ. Centre-ville, Montreal, Quebec, H3C 3J7 , Canada; • Teodor Gabriel Crainic; DIRO/Universite de Montreal, CRT, CP 7128 Succ Centre-Ville, Montreal, Quebec, H3C 3J7 , Canada; theo@crt.umontreal.ca The VRPTW is both of significant practical importance and difficult to solve from a mathematical programming point of view. We present a cooperative parallel metaheuristic strategy for the VRPTW built around a common memory (or pool or blackboard) that is fed by several constructive and improving heuristics, including several well-known tabu search approaches. SC20.3 Using Constraint-Based Operators to Solve the Vehicle Routing Problem with Time Windows • Michel Gendreau; DIRO/Universite de Montreal, CRT, CP 6128, succ. Centre-ville, Montreal, Quebec, H3C 3J7 , Canada; michelg@crt.umontreal.ca • Gilles Pesant; University of Montreal, CRT, CP 6128, Succ. Centre-ville, Montreal, Quebec, H3C 3J7 , Canada; • Louis-Martin Rousseau; University of Montreal, CRT, Montreal, Quebec, H3C 3J7 , Canada; We present operators searching large neighborhoods in order to solve the VRPTW. They make use of the pruning and propagation techniques of constraint programming which allow an efficient search of such neighborhoods. These operators are combined using a variable neighborhood descent search strategy. New best solutions are obtained on benchmarks. SC20.4 An Ant Colony System Heuristic for the Traveling Salesman Problem • Michel Gendreau; DIRO/Universite de Montreal, CRT, CP 6128, succ. Centre-ville, Montreal, Quebec, H3C 3J7 , Canada; michelg@crt.umontreal.ca • F.-X. LeLouarn; University of Montreal, CRT, CP 6128, succ. Centre-ville, Montreal, Quebec, H3C 3J7 , Canada; • Jean-Yves Potvin; University of Montreal, CRT, CP 6128, Succ, Centre-Ville, Montreal, Quebec, H3C 3J7 , Canada; The ACS is an optimization meta-heuristic based on the observation of ant colonies cooperation schemes. We present an application of ACS to the TSP in which the ants' action selection algorithm is the GENI procedure of Gendreau, Hertz & Laporte. Preliminary results reveal that the pheromone cues left by GENI ants do help other GENI ants to construct better TSP tours. # Session III Session: SC21 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: SICUP Room: Chair: Robert Haessler Chair Address: University of Michigan, Business Sch., Ann Arbor, MI 48109-1234 Chair E-mail: robert_haessler@ccmail.bus.umich.edu Chair: Chair Address: Chair E-mail: SC21.1 An Advanced Approach to the 2-Stage Cutting Stock Problem We propose a row and column generation technique for solving a 2-stage CSP. The technique is a generalization of the column generation method suggested by Gilmore & Gomory for solving a classic CSP. The procedure generates only those intermediate rolls and cutting patterns that are needed. SC21.2 The Cutting Stage in the Multi-Item Multi-Level Lot-Sizing Problem • Marcos Arenales; Universidade de Sao Paulo, Inst. de Ciencias Math. & Comp, Campus de Sao Carlos, CP 668, Sao Carlos, 13560-970 , Brazil; arenales@icmc.sc.usp.br • Maria Christina Nogueira Gramani; Universidade Estadual de Campinas, Fac. de Eng. Eletrica & Comp., CP 6101, Campinas, 13083-970 , Brazil; • Paulo M. Franca; Universidade Estadual de Campinas; The combined cutting stock and lot sizing problem consists of modeling a multistage production planning where the cutting stock problem is imbedded as the most important stage. In practice, usually these problems are heuristically solved separately. We formulated a mathematical model that solves both problems in SC21.3 Cutting Stock Problems: How to Select Stock Sizes Practical cutting stock situations both deal with short-term planning of orders with constant stock sizes as well as with long-term planning of stock sizes. Some typical situations are discussed. Some cases show how simple methods of stock size planning can reduce the losses of short term planning. SC21.4 no show • Yu Stoyan; Ukrainian National Academy of Science, Math Modeling & Optimal Design, Mech. Eng. Problems Inst., Kharkov, 610046 , Ukraine; stoyan@ipmach.kharkov.ua • G. Yaskov; Ukrainian National Academy of Science, Math Modeling & Optimal Design, Mech. Eng. Problems Inst., Kharkov, 610046 , Ukraine; # Roundtable: E-Commerce in Health Care Session: SC22 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Health Applications Section Track: Cluster: Room: Chair: William P. Pierskalla Chair Address: UCLA, The Anderson Sch., 110 Westwood Plaza, Ste. B411, Los Angeles, CA 90095-1481 Chair E-mail: william.pierskalla@anderson.ucla.edu Chair: Chair Address: Chair E-mail: SC22.1 Roundtable: E-Commerce in Health Care • Don Chenoweth; health2health.com, 650 South Cherry, Ste. 420, Denver, CO 80246-1806; dchenoweth@shs-denver.com • Liam O'Neill; Cornell University, Policy Analysis & Mgmt., N229 MVR Hall, Ithaca, NY 14853-4401; lo22@cornell.edu • Douglas A. Samuelson; InfoLogix, Inc., 8711 Chippendale Ct., Annandale, VA 22003; We examine the past, present, and future of e-commerce in health care. Special emphasis will be on business-to-business, HIPAA and network security. # Simulation & Optimization using Spreadsheets Session: SC23 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Simulation Section Track: Cluster: Room: Chair: John M. Charnes Chair Address: University of Kansas, Sch. of Business, 345 Summerfield Hall, Lawrence, KS 66045-2003 Chair E-mail: jmc@ukans.edu Chair: Chair Address: Chair E-mail: SC23.1 Introduction to Simulation & Optimization using Spreadsheets We will introduce the fundamentals of uncertainty analysis, Monte Carlo simulation and optimization modeling techniques using Crystal Ball 2000. An introductory model will be explored covering methodology for stochastic optimization of a budget constrained project selection. SC23.2 Simulation Models in the Classroom Crystal Ball has been used effectively in an introductory modeling course for MBA students. We discuss the cases, homework and classroom examples that illustrate how MBA student can be introduced to Monte Carlo simulation using Crystal Ball. SC23.3 Pricing American Put Options using Simulation & Optimization • John M. Charnes; University of Kansas, Sch. of Business, 345 Summerfield Hall, Lawrence, KS 66045-2003; jmc@ukans.edu Crystal Ball 2000 and OptQuest can be used to compute values for American put options. We will demonstrate the modeling technique and weigh its advantages and disadvantages. # Virtual Environments for Manufacturing Systems Design Session: SC24 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Material Handling Room: Chair: Pat Banerjee Chair Address: University of Illinois, Ind. Virtual Reality Inst., Dept. of Mech. Eng., Chicago, IL Chair E-mail: Chair: Chair Address: Chair E-mail: SC24.1 An Intelligent Human-Object Interaction Environment for Rapid Manufacturing Systems Design • A. Banerjee; University of Texas; • Pat Banerjee; University of Illinois, Ind. Virtual Reality Inst., Dept. of Mech. Eng., Chicago, IL; A collaborative virtual environment is interfaced to a simulation engine for driving the system and a force feedback control system for effective interaction of humans with the environment. Some aspects of optimal object data encapsulation for reducing data latency are addressed here. SC24.2 Simulation of a Software Architecture for a Mobile Robotic-Based Shop Floor Design • T. Kesavadas; University of Buffalo, Virtual Reality Lab., Buffalo, NY 14260; • Yuri Menezes; University of Buffalo, Virtual Reality Lab., Buffalo, NY 14260; We will present a new software architecture for a proposed factory layout scheme based on continuously self rearranging robotic cells. A simulation-based interface will be used to create and analyze such an environment. Initial results will be presented. SC24.3 Tele-Collaborative Simulation between Existing Manufacturing Systems & Virtual Reality Replicas • Ali Akgunduz; University of Illinois, Ind. Virtual Reality Inst., Chicago, IL; • Pat Banerjee; University of Illinois, Ind. Virtual Reality Inst., Dept. of Mech. Eng., Chicago, IL; We describe a real-time system simulation model for replacing equipment in existing manufacturing facilities using tele-collaborative VR models. For equipment selection during the replacement process, a number of criteria, i.e., size, compatibility with existing machineries, compatibility with existing products, energy usage, etc., are investigated. # Nonlinear Programming & Applications Session: SC25 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Optimization Section Track: Cluster: Nonlinear Programming Room: Chair: Mihai Anitescu Chair Address: University of Pittsburgh, Dept. of Math, Thackery 301, Pittsburgh, PA 15260 Chair E-mail: anitescu@math.pitt.edu Chair: Chair Address: Chair E-mail: SC25.1 Quasi-Newton Interior Algorithms for Nonlinear Programming • Michael Wagner; Old Dominion University, Dept. of Math & Stats., 500 Batten Arts & Letters, Norfolk, VA 23529-0077; wagner@orie.cornell.edu Quasi-Newton methods for unconstrained problems have been around for over 40 years now and more recently they have been applied to equality-constrained problems. We investigate ways to extend some of these approaches to the more general case with inequality constraints. SC25.2 Degenerate Nonlinear Programming • Mihai Anitescu; University of Pittsburgh, Dept. of Math, Thackery 301, Pittsburgh, PA 15260; anitescu@math.pitt.edu We investigate nonlinear programming problems for which the gradients of the active constraints are linearly or nearly linearly independent. We discuss the behavior of sequential quadratic programming algorithms when applied to such problems. SC25.3 Solving Optimal Control Problems using SNOPT & Collocation • Michael Gertz; Argonne National Laboratory, MCS Bldg. 221, Argonne, IL 60439; gertz@ncs.anl.gov • Philip E. Gill; University of California, Dept. of Mathematics, 9500 Gilman Dr., MC 0112, La Jolla, CA 92093-0112; pgill@ucsd.edu • Julia Muetherig; University of California, Dept. of Mathematics, 9500 Gilman Dr., MC 0112, La Jolla, CA 92093-0112; julia@sdna0.ucsd.edu • J. Ben Rosen; University of California, Dept. of Mathematics, 9500 Gilman Dr., MC 0112, La Jolla, CA 92093-0112; jbrosen@cs.uscsd.edu Collocation techniques provide a set of rules for transforming an optimal control problem to an NLP. In practice, coding these problems correctly is difficult and time consuming. We present software that we have developed for automating the transformation and solving the problem using SNOPT. SC25.4 COOPT: A Software Package for Optimal Control of Large-Scale Differential-Algebraic Equation Systems • Radu Serban; University of California, Dept. of Mech. & Environ. Eng., Santa Barbara, CA 93106; radu@engineering.ucsb.edu • Linda R. Petzold; University of California, Dept. of Mech. & Environ. Eng., Santa Barbara, CA 93106; petzold@engineering.ucsb.edu • Shengtai Li; University of California, Dept. of Mech. & Environ. Eng., Santa Barbara, CA 93106; shengtai@cs.ucsb.edu This package implements a direct method with modified multiple shooting type techniques for solving optimal control problems of large-scale DAE systems. The resulting optimization problem is solved by sparse SQP methods. COOPT has been used in chemical vapor deposition of superconducting thin films, spacecraft trajectory design and contingency/recovery problems and computation of cell traction forces in tissue engineering. # Telecommunication Systems Session: SC26 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Stochastic Models & Applications Room: Chair: Emmanuel Fernandez Chair Address: University of Arizona, SIE Dept., Tucson, AZ 85721-0020 Chair E-mail: emmanuel@sie.arizona.edu Chair: Chair Address: Chair E-mail: SC26.1 Information Technology & Telecommunications in the University of Texas System The University of Texas System has embarked in a far reaching plan to develop and enhance the telecommunication infrastructure and information technology throughout the system. We present an overview of objectives and programs, e.g., the UT Telecampus. SC26.2 A Markov Decision Model for Intruder Location in IP Networks • Mark A. Shayman; University of Maryland, ECE Dept., College Park, MD 20742; shayman@eng.umd.edu • T. Darling; University of Maryland, Dept. of ECE, College Park, MD 20742; The problem of locating an intruder or misuser in an autonomous system in the Internet can be formulated as a Markov decision problem whose state evolves on a set of trees. For networks of realistic size, exact calculation of optimal policies by dynamic programming is not feasible. Approximation techniques such as neurodynamic programming are investigated. SC26.3 Fault Management in Communication Networks: Test Scheduling with a Risk-Sensitive Criterion & Precedence Constraints We consider the problem of determining the optimal sequence of tests for the discovery of a faulty component in a telecommunications network, where there is a random cost associated with testing a component. A risk-sensitive performance criterion is used in order to rank different competing schedules. # Tutorial: Breaking the Mold: Using Commercial Software & Experiential Learning to Teach the First MS Course Session: SC27 Date/Time: Sunday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Room: Chair: Matthew J. Liberatore Chair Address: Villanova University, Dept. of Decision & IT, Villanova, PA 19085 Chair E-mail: matthew.liberatore@villanova.edu Chair: Robert L. Nydick Chair Address: Villanova University, Dept. of Decision & IT, Villanova, PA 19085 Chair E-mail: robert.nydick@villanova.edu SC27.1 Tutorial: Breaking the Mold: Using Commercial Software & Experiential Learning to Teach the First MS Course Our approach to teaching the first course in MS is predicated upon 2 beliefs: MS is more important to business education than ever and students can implement MS more easily than ever. What is required is a radical change in the way the first MS course is conducted. Spreadsheets offer important advantages in teaching MS. However, we offer a proven alternative to spreadsheets... # What Defines the Success of OR Applications in the Railroad Industry I Session: SC28 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: RASIG Track: Cluster: Room: Chair: Ingrid Schultze Chair Address: Reebie Associates Chair E-mail: ischultze@reebie.com Chair: Chair Address: Chair E-mail: SC28.1 What Defines the Success of OR Applications in the Railroad Industry I • Carl D. Martland; MIT, Dept. of Civil & Environ. Eng., Rm. 1-153, Cambridge, MA 02139; • Bengt Muten; Reebie Associates; • Jason Kuehn; Multimodal Applied Systems, Inc.; • Roger W. Baugher; Norfolk Southern Railroad; rwbaughe@nscorp.com • Chip Kraft; AMTRAK; • Howard A. Rosen; ALK Associates; • Steve Ditmeyer; Federal Rail Administration; Overview & Introduction: Carl MartlandA - Network & Strategic Planning ApplicationsIngrid Schultze: ModeratorConsultant Perspective - Bengt Muten: Successes/failures of network planning & diversion models.Industry Perspective - Jason Kuehn: Contributions of OR to industry-wide restructuring efforts such as M&A, mixing centers, etc., where OR has worked, where it 'needs work.'Systems Perspective - Roger Baugher: The system perspective - how technology advances are shaping the application of OR in the rail industry.B - Corridor-Level Planning ApplicationsChip Kraft: Moderator - Overview of How We MeasureConsultant Perspective - Howard Rosen: Successes and failures of corridor planning models from a systems and consulting perspective. Industry - Steve Ditmeyer Systems Perspective - Roger Baugher: Contributions of OR to corridor planning efforts - where OR has worked, where it needs work. # Modeling Price Formation & Market Power in Competitive Electricity Markets II Session: SC29 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: ENRE Track: Cluster: Room: Chair: James B. Bushnell Chair Address: University of California, Energy Inst., 2539 Channing Way, Berkeley, CA 94720-5180, USA Chair E-mail: jimb@ieor.berkeley.edu Chair: Chair Address: Chair E-mail: SC29.1 Competition in Markets for Electricity: A Conjectured Supply Function Approach Competition between electricity suppliers who form conjectures about opponents' supply functions is formulated as a linear complementarity problem. Transmission is constrained by a DC network. The model has the advantage of being able to obtain equilibria even when demand is perfectly inelastic. The model is applied to the UK system. SC29.2 The Effects of Technology Segmentation Market Power in Wholesale Electricity Markets • John Bower; London Business School, Sussex Place, Regent's Park, London, NW1 4SA , England, UK; jbower@london.edu • Derek W. Bunn; London Business School, Sussex Place, Regents Park, London, NW1 4SA , England, UK; dbunn@london.edu Traditional analysis of market power in deregulated electricity markets concentrates on the impact of industry concentration and market mechanisms. Using an agent-based simulation model, we show that the allocation of generating technology between firms in a national electricity industry can also have significant impact. Simulation results are compared with empirical observations from the German and UK power markets. SC29.3 The Extent of the Electricity Market • Thomas J. Overbye; University of Illinois, Urbana, IL; overbye@powerworld.com • Douglas Hale; US Department of Energy, Energy Info. Admin., 1000 Independence Ave. SW, Washington, DC 20585; douglas.r.hale@eia.doe.gov • Thomas Leckey; PowerWorld Corp., Urbana, IL; A model of the Eastern Interconnect was constructed using (FERC) Form 715 filings to provide a detailed representation of the transmission system and FERC Form 1 data with information from the EIA National Energy Modeling System to represent generator costs. An optimal power flow was then used to optimally dispatch the East both under administered and free trade conditions... SC29.4 Price Convergence in California's Wholesale Electricity Markets • Severin Borenstein; University of California, Energy Inst., 2359 Channing Way, Berkeley, CA 94720; borenste@haas.berkeley.edu • James B. Bushnell; University of California, Energy Inst., 2539 Channing Way, Berkeley, CA 94720-5180, USA; jimb@ieor.berkeley.edu • Chris Knittel; Boston University, Dept. of Finance & Econ., 595 Commonwealth Ave., Boston, MA 02215; knittel@bu.edu • Catherine Wolfram; Harvard University, Dept. of Economics, Cambridge, MA 02138-3001; We study the relationships among prices in different markets within the California electricity industry during the last year. In the absence of institutional impediments, economic forces will lead to the convergence of prices in different markets for delivery of power at the same time. Many institutional impediments did exist, however. # Web-Based Tactical Applications of OR I Session: SC30 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Computing Society Track: Cluster: Room: Chair: Radhika Kulkarni Chair Address: SAS Institute Inc., SAS Campus Dr., Cary, NC 27513 Chair E-mail: radhika.kulkarni@sas.com Chair: Chair Address: Chair E-mail: SC30.1 Web Technologies for Decision Support • Robert Fourer; Northwestern University, Dept. of IEMS, 2145 Sheridan Rd., Tech. Inst., Evanston, IL 60208; 4er@iems.nwu.edu Software developers of all kinds are becoming 'application service providers' by converting their conventional software packages to web-based services. We survey web technologies that are (or will soon be) available to support the choice or use of decision support software, with examples from diverse optimization services and systems. SC30.2 Web-Enabled Inventory Replenishment Planning We discuss development of a Web-enabled software application for inventory replenishment planning. Using a distributed architecture, we access historical demand data to calculate continuous review (s, S) policies with service level constraints for different classes of products. Policy results, performance analysis and data statistics are made available via a Web browser. E-mail transmission of product orders is supported. SC30.3 Using Optimization for Smart Internet Marketing Decisions • Harlan Crowder; Hewlett-Packard Laboratories, 1501 Page Mill Rd., Palo Alto, CA 94304; harlan_crowder@hp.com Building, executing and evaluating marketing campaigns was once a leisurely process that unfolded over weeks or months. The intrinsic response time associated with television and direct mail campaigns allowed marketers the liberty of unhurried decisions. Then came the Internet. We will describe an experimental system for quickly designing, testing and optimizing on-line marketing campaigns. # Tutorial: XPRESS-MP: A Fully Integrated Optimization System Session: SC31 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Computing Society Track: Cluster: Room: Chair: Alkis Vazacopoulos Chair Address: Dash Optimization, Inc., 135 West 27th St., 7th Fl., New York, NY 10001 Chair E-mail: alkis@dashopt.com Chair: Chair Address: Chair E-mail: SC31.1 Tutorial: XPRESS-MP: A Fully Integrated Optimization System • Alkis Vazacopoulos; Dash Optimization, Inc., 135 West 27th St., 7th Fl., New York, NY 10001; alkis@dashopt.com XPRESS-MP is a state-of-the-art software solution for modeling and optimization. We will show how XPRESS-MP can be used to model and solve complex problems arising in a wide variety of applications. We will introduce our new product XBSL which enables you to build up models dynamically within a high level programming language... # Evolutionary Perspectives on Firm Strategy Session: SC32 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Organization Science Section Track: Cluster: Room: Chair: Peter W. Roberts Chair Address: Carnegie Mellon University, GSIA, 5000 Forbes Ave., Pittsburgh, PA 15213-3890 Chair E-mail: proberts@andrew.cmu.edu Chair: Chair Address: Chair E-mail: SC32.1 Incumbent Resurgence: Strategies for Surviving Competence-Destroying Environmental Shocks • Raja Roy; University of Pittsburgh, Katx. Grad. Sch. of Bus., 247 Mervis Hall, Pittsburgh, PA 15260; • Susan K. McEvily; University of Pittsburgh, Katz Grad. Sch. of Bus., 247 Mervis Hall, Pittsburgh, PA 15260; smcevily@katz.pitt.edu Following radical technological change, incumbents pursue strategies that either enhance their survival prospects or their odds of regaining competitive advantage. Data from the machine tool industry supports this proposition, but suggests firms manage the tradeoff by acquiring rather than developing new capabilities. Implications for the evolution of competence are discussed. SC32.2 Choice Interaction & Organizational Structure • Jan Rivkin; Harvard Business School, 239 Morgan Hall, Boston, MA 02163; jrivkin@hbs.edu • Nicolaj Siggelkow; University of Pennsylvania, The Wharton Sch., 2017 SH-DH, Philadelphia, PA 19104; We use an agent-based simulation to examine how organizations should allocate interdependent decisions to independent managers. We explore, inter alia, the importance of assigning related decisions to a single manager, the robustness of organizational designs to management mistakes and the evolution of organizational structures to mimic underlying interactions among choices. SC32.3 Firms as Portfolios of Strategic Attributes: An Austrian Economic Approach to Financial Performance Heterogeneity • Peter W. Roberts; Carnegie Mellon University, GSIA, 5000 Forbes Ave., Pittsburgh, PA 15213-3890; proberts@andrew.cmu.edu We develop an Austrian explanation of the dynamic heterogeneity in profitability from 3 premises: firms are systems of strategic attributes, attributes have direct and interactive effects on profitability and managers always strive to improve profitability. Our explanation highlights Kirzner's opportunity-driven discovery, Lachmann's heterogeneous capital resources and Hayek's emergent, decentralized knowledge. SC32.4 Making E-Commerce Strategies in Established Firms • Gabriel Szulanski; University of Pennsylvania, The Wharton Sch., 2033 SH-DH, Philadelphia, PA 19104-6370; szulanski@wharton.upenn.edu I explore the process of creation of e-commerce strategy in established firms. While some attention has been devoted to e-commerce strategy, the organizational processes that lead to those strategies have not been studied closely. I use an evolutionary framework to compare and contrast strategy-making processes within 3 established companies. # Production Planning Session: SC33 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Funda Sahin Chair Address: Texas A&M University, Grad. Sch. of Bus., Dept. of IOM, College Station, TX 77843-4217 Chair E-mail: fsahin@cgsb.tamu.edu Chair: Chair Address: Chair E-mail: SC33.1 Analyzing the Effects of Distillate Fuel Oil Production on the Heating Oil Futures Contract • James R. Porter; Cornell University, Dept. of OR/IE, Rhodes Hall, 2nd Fl., Ithaca, NY 14853; porter@orie.cornell.edu • Robin Roundy; Cornell University, Dept. of OR/IE, 216 Rhodes Hall, Ithaca, NY 14853; robin@orie.cornell.edu Our research is concerned with analyzing the effects of production decisions on the price of the heating oil futures contract. A stochastic dynamic programming problem is formulated to minimize refinery cost while matching past production decisions. The solution is then perturbed in an effort to value future units of inventory. SC33.2 A Cyclic Scheduling Problem with Tool Transportation Times We consider a cyclic 2-machine scheduling problem with tool-sharing and positive tool transportation times. For finding an optimal schedule, only schedules of a few different structures need to be taken into account. With this knowledge, a fast algorithm is obtained. SC33.3 no show SC33.4 Economic Production Lot-Size with Periodic Fixed Costs & Overtime • Funda Sahin; Texas A&M University, Grad. Sch. of Bus., Dept. of IOM, College Station, TX 77843-4217; fsahin@cgsb.tamu.edu • Powell Robinson; Texas A&M University, Grad. Sch. of Bus., Dept. of IOM, College Station, TX 77843-4217; probinson@cgsb.tamu.edu We propose mathematical formulations and algorithms that consider equipment setup, periodic fixed costs, regular and overtime production constraints. We documented the performance of the algorithms on a set of simulated and actual test problems. The potential economic benefit of the procedures is promising. # Decision Analysis III Session: SC34 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: David Taylor Chair Address: Capital One, 5600 Cox Rd., MS 12011-0330, Glen Allen, VA 23060 Chair E-mail: dave.taylor@capitalone.com Chair: Chair Address: Chair E-mail: SC34.1 Practical Decision Analysis Tools IT project selection and resource allocation was painfully inefficient until the IT analysis team introduced practical DA into the corporate culture. This team introduced value focused thinking in all of Capital One's strategic business unit, demonstrating the ability to quantify the value of ideas in other than economic terms. SC34.2 Process Change to Enable a Decision Analysis Culture In the entrepreneurial environment that defines Capital One's culture, introducing quantitative decision analysis techniques required a robust, yet flexible process. We describe how the IT analysis team adeptly melded value focused thinking and linear programming into a transparent and efficient process. SC34.3 Adapting Off-the-Shelf Decision Analysis Software Although off-the-shelf DA tools are invaluable for decision analysts, they are rarely suitable for business users. We discuss customizations made to existing software to enable Capital One business users to utilize value focused thinking through an interactive web-based GUI to build business cases and make rational decisions. SC34.4 Client & Toolmaker Partnerships Two trends are changing how DA vendors work with clients: growth of DA expertise within corporations and use of internet/intranet to strategically fuse DA methodologies into the corporation's backbone. InfoHarvest and Capital One are collaborating to realize Capital One's vision of making value focused thinking a pervasive IT service. # Information Systems I Session: SC35 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Parviz Ghandforoush Chair Address: Virginia Tech., 7054 Haycock Rd., Falls Church, VA 22043 Chair E-mail: pghandfo@vt.edu Chair: Chair Address: Chair E-mail: SC35.1 Business Process Reengineering/Information Systems & Competitive Advantages Today, globalization is inevitable and organizations must cope with this new business environment which requires a new form of organization. Management attention has moved from competitive advantage to strategic competitive survival. Quality, innovation and customer service are now more important to survival than cost, growth and control... SC35.2 The Influence of Software Process Maturity & IS Management Maturity on the Attitude Adopted in the IS Requirements Definition Activity A set of case studies brings confidence to the belief that organizations with a higher level of maturity, i.e., software process maturity and information systems management function maturity, tend to treat the definition of IS requirements as an organizational activity and not just as a technical (information technology related) activity. SC35.3 Infusion of Enterprise Resource Planning Systems • Paul D. Brown; Georgia State University, 35 Broad St., Atlanta, GA 30303; brownpries@aol.com An important new technology of the 1990s has been ERP systems. ERP systems have been a key factor in process reengineering, supply-chain partnering, management restructuring and other strategic initiatives. Our objective is to derive a model that explains the level of infusion of ERP systems within organizations. SC35.4 Design of an E-Business Portal for Transportation Planning • Parviz Ghandforoush; Virginia Tech., 7054 Haycock Rd., Falls Church, VA 22043; pghandfo@vt.edu • Tarun Sen; Virginia Tech., 7045 Haycock Rd., Falls Church, VA 44053; tksen@vt.edu We describe a prototype of a portal for transportation businesses that incorporates routing and planning algorithms. Among the primary objectives of an e-business portal is to attract customers to the portal to induce revenue generation. There are several revenue generation models. A particular revenue generation model will be adopted for this study, i.e., a fee for service... # Logistics Management I Session: SC36 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Atsuo Suzuki Chair Address: Nanzan University, Dept. of Math. Sci., 27 Seirei-cho, Seto-shi, Aichi, 489-0863 , Japan Chair E-mail: atsuo@ms.nanzan-u.ac.jp Chair: Chair Address: Chair E-mail: SC36.1 The Two-Carousel Storage Location Problem We describe a storage location problem in a 2-carousel storage and retrieval system. We present an analytical model for the problem and discuss some heuristic solution procedures. SC36.2 A Hierarchical Fleet Planning Framework for the Car Rental Industry • Julian E. Pachon; Caleb Technologies Corp., 9130 Jollyville Rd., Ste. 100, Austin, TX 78759; jpachon@coeds.eng.miami.edu • Lefteris T. Iakovou; University of Miami, Dept. of IEN, PO Box 248294, Coral Gables, FL 33124; eiakovou@miami.edu • Chi M. Ip; ANC Rental Corp., 110 SE 6th St., Ft. Lauderdale, FL 33301; ipc@goalamo.com We present the development of a novel hierarchical fleet planning framework that synthesizes all 3 major phases of the decision-making process that car rental companies utilize to maximize their yield. We develop appropriate solution methodologies and exhibit their application via a real case study for the state of Florida. SC36.3 withdrawn - author request of 10/13 SC36.4 A Rental Fleet Sizing Model & a Two-Phase Algorithm An integrated rental fleet sizing model is formulated. A demand-shifting feasibility algorithm is developed for solving Benders subproblems and for obtaining Lagrangean feasible solutions efficiently. A 2-phase algorithm and an end-effect correction of infinite horizon are further discussed with numerical experiments. SC36.5 A Continuous Hub Location Problem on the Sphere • Atsuo Suzuki; Nanzan University, Dept. of Math. Sci., 27 Seirei-cho, Seto-shi, Aichi, 489-0863 , Japan; atsuo@ms.nanzan-u.ac.jp We formulate a hub location problem on the sphere using the sphere Voronoi diagram. Assume that the airports are spread continuously on the sphere, the airports connect to the nearest hub airport and the route of travelers has 2 hub stops. Then, the territory of the hubs are their Voronoi region on the sphere. We show the optimal layout of hubs when the number of hubs n is small, and suboptimallayouts when n is large. # Optimization Techniques III Session: SC37 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Dipak Chaudhuri Chair Address: Tulane University, Freeman Sch. of Bus., 7 McAlister Dr., New Orleans, LA 70118 Chair E-mail: dipak.chaudhuri@tulane.edu,, http://www.tulane.edu/~dipak_c/dipak.html Chair: Chair Address: Chair E-mail: SC37.1 Applying Real Option Theory to Engineering Problems We apply Merton's optimality conditions and Ito's Lemma (Dixit, 1994) for developing the mathematical tools needed to solve optimal control problems in batch distillation columns operating under uncertainty. A qualitative analysis is presented and some numericalexperiments are used to show the scope and usefulness of such an analysis. SC37.2 Analysis of the Affine Scaling Method for Semidefinite Programming • Beong Choi; University of Michigan, IOE Dept., 1205 Beal Ave., Ann Arbor, MI 48109-2117; bchoi@umich.edu • Romesh Saigal; University of Michigan, IOE Dept., 1205 Beal Ave., Ann Arbor, MI 48109-2117; rsaigal@umich.edu The affine scaling method for linear programming problems has been well studied and its convergence proven. However, its straightforward extension to SDP problems has been shown to fail. We present a slightly modified approach to the primal affine scaling method for SDPs. Numerical experiments show that the algorithm works well on various types of problems... SC37.3 Java Constrainer: A Constraint Satisfaction Environment for Web Applications Java Constrainer is a generic constraint solver for Java developed by IntelEngine, Inc. It demonstrates the efficiency and expressiveness comparable with C++ ILOG Solver. We will describe the major functionality, implementation issues and the comparison results using well-known benchmark problems. SC37.4 Inspection Schedules & Minimal Repairs for Optimal Profit per Cycle of Deteriorating Production Systems Production systems need to be inspected at predetermined intervals to detect presence of any non self-announcing failure of components. If such failures are detected at any inspection point, a minimal repair is performed on the system by replacing the failed components and the system is restarted after scrapping the entire production during the interval since the last inspection... # Industry Applications I Session: SC38 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: K. Jo Min Chair Address: Iowa State University, Dept. of IMSE, 2019 Black, Ames, IA 50011 Chair E-mail: jomin@iastate.edu Chair: Chair Address: Chair E-mail: SC38.1 SAP Implementation in a Major Steel Plant in India • Rajiv R. Singh; K. M. Gupta & Co., Ispat Industries Ltd., A-10/1 MIDC Industrial Area, Kalmeshwar, 441501 , India; rajiv_online@yahoo.com We present lessons from SAP implementation in a major steel processing plant in central India. Real-time information about enterprise resources and constraints have presented unprecedented opportunities for operational improvements. Yet, management culture, more than technology, seems to determine the outcome. SC38.2 An Optimization Model for the Strategic & Economic Development of a Petrochemical Industry Today, there is a shift in strategic thinking arising from the increasing importance of technology and innovation to the overall strategy of the industry. These 2 aspects of the industry's activity can no longer be treated as a tactical issue. The resource development and environmental interactions are increasingly dependent on the availability of innovative products, processes and services... SC38.3 No Title Supplied • Boo-Sik Kang; Korea Telecom, Telecom Network Lab., 463-1 Junmin-dong Yusong-gu, Taejon, 305-300 , Korea; richgang@kt.co.kr • Byoong-Wook Lee; Korea Telecom, Telecom Network Lab., 463-1 Junmin-dong Yusong-gu, Taejon, 305-300 , Korea; leebw@kt.co.kr • Young-Jin Sim; Korea Telecom, Telecom Network Lab., 463-1 Junmin-dong Yusong-gu, Taejon, 305-300 , Korea; yjsim@kt.co.kr Troubles of data communications services must be handled by interconnection tests among various systems. We propose a model used machine learning techniques to search site and causes of the troubles by setting intitial test strategy with the trouble type and using prior test results to determine posterior test methods. SC38.4 A Real Options Model for Multiple Inter-Related Generation Projects • K. Jo Min; Iowa State University, Dept. of IMSE, 2019 Black, Ames, IA 50011; jomin@iastate.edu • Chung-Hsiao Wang; Iowa State University, Dept. of IMSE, 2019 Black, Ames, IA 50011; wchrist@iastate.edu The electric power generation business is moving toward uncertain and competitive environments. We design and analyze a multi-unit generation planning model based on the real options theory. Under the assumption of inter-related generation unit values, we derive various managerial insights and economic implications via a numerical example. # Marketing III Session: SC39 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Timothy L. Urban Chair Address: University of Tulsa, 600 South College Ave., Tulsa, OK 74012 Chair E-mail: urbantl@utulsa.edu,, http://www.cba.utulsa.edu/qm_tlu/ Chair: Chair Address: Chair E-mail: SC39.1 withdrawn - author request of 10/30 SC39.2 A Market Structure Model in Consumer Durable We focus on the identification of competitive market structure using brand-switching data in consumer durable. We propose a new approach that generates switching constants and asymmetric proximity data. Pictorial representations will be drawn using MDS and hierarchical clustering analysis. Finally, we compare our results with those of previous researches. SC39.3 Shelf Management & Product Placement It is a well-known and empirically-verified fact that a product's shelf location (its vertical and horizontal positioning) has a significant impact on sales for many retail items. Motivated by the operations of a retailer in the highly-competitive beverage market, we present a comprehensive shelf-management model that incorporates product placement, with consideration given to maintaining the grouping of product families. SC39.4 A Scoring Model to Optimize Direct-Mail Marketing Promotions for a Pharmacy Benefit Manager • Edward S. Binkowski; Data Analysis Services, New York, NY; • Moshe B. Rosenwein; Merck-Medco, Franklin Lakes, NJ; moshe_rosenwein@merck.com We describe a scoring model that supports direct mail promotions, targeting potential e-commerce customers for a major pharmacy benefit manager (PBM). Advanced data mining methods are applied across critical PBM data marts to determine attributes that influence customer behavior. The model may potentially reduce direct mailing costs by 40%. SC39.5 withdrawn - author request of 10/6 • Eric C. Jackson; Michigan State University, Eli Broad Coll. of Bus., East Lansing, MI 48824; jacks351@pilot.msu.edu • Ram Narasimhan; Michigan State University, Eli Broad Coll. of Bus., East Lansing, MI 48824; narasimh@msu.edu • David Mendez; University of Michigan, Health Mgmt. & Policy, Ann Arbor, MI 48109; dmendez@umich.edu # Statistics & Quality Control Session: SC40 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: M. Jeya Chandra Chair Address: Pennsylvania State University, 310 Leonhard Bldg., University Park, PA 16802 Chair E-mail: mjc3@psu.edu Chair: Chair Address: Chair E-mail: SC40.1 Multivariate Capability Indices using Process-Oriented Basis Representations Negotiations between a manufacturer and a parts supplier often involve the length of time the manufacturer is contractually held to its order quantity. We find conditions under which the manufacturer is better off with a contract that requires an early commitment to its order quantity, before the supplier commits resources. SC40.2 SPC for Multivariate Batch Processes with Mid-Bath Intervention • Robin C. Wurl; Rutgers University, 96 Frelinghuysen Rd., Piscataway, NJ 08854; wurl@rci.rutgers.edu • Susan L. Albin; Rutgers University, Dept. of IE, 96 Frelinghuysen Rd., Piscataway, NJ 08854; salbin@rci.rutgers.edu Consider multivariate batch processes where one can intervene during each batch and there is significant time between batches. We construct control charts to monitor within-batch and the batch-to-batch variation, taking into account autocorrelations within a batch. We use PLS models and bootstrap techniques and apply the method to filament extrusion. SC40.3 A Weibull Regression Model using Nomimal-the-Best Quality Characteristics as Covariates • M. Jeya Chandra; Pennsylvania State University, 310 Leonhard Bldg., University Park, PA 16802; mjc3@psu.edu • Murray M. Smith; University of Auckland, Private Bag 92019, Auckland, , New Zealand; mh.smith@auckland.ac.nz A model to capture the effects of quality characteristics on the lifetime of a product is formulated. The lifetime distribution of an individual item, conditional on the values of certain quality characteristics, is Weibull. The rate parameter is a function of the capability indices of the quality characteristics. SC40.4 Constructing a Double Sampling X-Bar Control Chart for Agile Manufacturing using a Genetic Algorithm • David W. He; University of Illinois, Dept. of Mech. Eng., MC 251, 842 West Taylor St., Rm. 3049, Chicago, IL 60607-7022; davidhe@uic.edu Double sampling (DD) x-bar control chart is designed to allow quick detection of a small shift of process mean and provides a quick response in an agile manufacturing environment. A new formulation for constructing an optimal DD x-bar control chart is developed and solved with a GA. SC40.5 Process Investment & Loss Functions: Models & Analysis • Shailesh S. Kulkarni; University of North Texas, BCIS Dept., Box 305249, Denton, TX 76203-5249; kulkarni@unt.edu • Victor R. Prybutok; University of North Texas, BCIS Dept., Box 305249, Denton, TX 76203-5249; prybutok@unt.edu Recent studies have looked at process variance reduction from a non-traditional Taguchi quality cost perspective. We propose an alternative to Taguchi variance reduction models. We develop analytical models and investigate structural properties. We derive optimal investment levels for reduction in process variance and provide several practical insights. # Tutorial: The Roundtable Presents - Small Business Innovative Research Grants & How to Win Them Session: SC41 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: INFORMS Roundtable Track: Cluster: Room: Chair: Joseph H. Discenza Chair Address: SmartCrane, Inc., a Daniel H. Wagner Associates Company, 2 Eaton St., Ste. 500, Hampton, VA 23669 Chair E-mail: joeh@discenza.com Chair: Chair Address: Chair E-mail: SC41.1 Tutorial: The Roundtable Presents Small Business Innovative Research Grants & How to Win Them • Joseph H. Discenza; SmartCrane, Inc., a Daniel H. Wagner Associates Company, 2 Eaton St., Ste. 500, Hampton, VA 23669; joeh@discenza.com Small business innovative research programs provide opportunities for practitioners, new graduates and university departments to obtain seed funding for investigation of new methods and development of new commercial products. This tutorial will cover topic selection, proposal writing and strategies for commercialization. # No Title Supplied Session: SC42 Date/Time: Sunday 13:15-14:45 Type: Sponsored Sponsor: Technology Management Section Track: Cluster: Room: Chair: Victoria L. Mitchell Chair Address: Ohio State University, Acct. & MIS, Fisher Coll. Bus., 2100 Neil Ave., Columbus, OH 43210 Chair E-mail: mitchell@cob.ohio-state.edu Chair: Chair Address: Chair E-mail: SC42.1 Technology Business Incubation in North America: Prospects for Cooperation in Regional Business Development • Sarfraz A. Mian; SUNY, Sch. of Business, Oswego, NY 13126; • Leonel Corona; UNAM, , , Mexico; • Jerome Doutriaux; University of Ottawa, , , Canada; We present preliminary results of our recent survey of selected technology business incubation facilities from the NAFTA region. Focusing on tenant entrepreneurs from these facilities, the data was collected from multiple university, private sector and government stakeholders. The study reveals some unique characteristics of the incubated firms and their nurturing environment, providing insights for the existing and new entrepreneurs... SC42.2 Architectural Competence & Business Platform Change: A Knowledge Perspective • Victoria L. Mitchell; Ohio State University, Acct. & MIS, Fisher Coll. Bus., 2100 Neil Ave., Columbus, OH 43210; mitchell@cob.ohio-state.edu Architectural knowledge refers to an understanding of how the components of a system are integrated. Architectural competence refers to an organization's ability to manage architectural knowledge. We refine the measure for architectural competence and measure its impact on evolving business platforms. # Risk Management III Session: SC43 Date/Time: Sunday 13:15-14:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Reinhart Schmidt Chair Address: University of Halle, Dept. of Finance & Banking, Universitatsplatz 8, Halle, D-06099 , Germany Chair E-mail: reinhart.schmidt@wiwi.uni-halle.de Chair: Chair Address: Chair E-mail: SC43.1 no show • Silvio R. B. de Gouvea; Billiton Metais SA, Praia de Botafogo 228, 4th Fl., Rio De Janeiro, 22359-900 , Brazil; silvio@billiton.com.br • David Benarroch; Billiton Metais SA, Praia de Botafogo 228, 4th Floor, Rio de Janeiro, 22359-900 , Brazil; SC43.2 Managing Electricity Reliability Risk through the Futures Markets In order to characterize the behavior of market participants in a competitive electricity market with reliability requirements, we model spot and futures markets for both electricity and ancillaryservices. We evaluate the optimal level of trading as well as the equilibrium prices in each market. SC43.3 Shareholder Value-at-Risk: Valuation of Corporate Strategies under Chance-Constraints A multiperiod corporate planning model under risk is developed for the evaluation of a firm's value for alternative sets of strategies. Using stochastic simulation, the probability distribution of the firm value and other probabilities are derived. Individual investors' preferences are used to compute shareholder values. # Software Demonstration I Session: SC45 Date/Time: Sunday 13:15-14:45 Type: Software Demo Sponsor: Track: Cluster: Room: Chair: Sam L. Savage Chair Address: Stanford University, Dept. of MSE, 417 Terman Eng. Ctr., Stanford, CA 94305-4023 Chair E-mail: savage@stanford.edu Chair: James Fitzsimmons Chair Address: McGraw-Hill/Irwin Chair E-mail: jfitz@mail.utexas.edu SC45.1 INSIGHT.xla: Business Analysis Software • Sam L. Savage; Stanford University, Dept. of MSE, 417 Terman Eng. Ctr., Stanford, CA 94305-4023; savage@stanford.edu INSIGHT.xla: Business Analysis Software for Microsoft Excel is a suite of Excel add-ins for simulation, decision analysis, forecasting, queueing and optimization. Sam Savage will demonstrate how easy it is to use INSIGHT with practical illustrative examples and discuss how INSIGHT with its accompanying textbook can be used in the classroom. SC45.2 Effective Use of Simulations in Teaching Service Operations Management The attributes and advantages of ServiceModel, the professional simulation software from the ProModel Corporation, will be addressed along with how Professor Fitzsimmons has integrated its use into his course on service management. As time permits, we will also demonstrate various electronic solutions, available from McGraw-Hill, to implement online courses. # COIN Installfest for Windows: A Hands-On Workshop using Open Source Software for OR Session: SC46 Date/Time: Sunday 13:15-14:45 Type: Sponsor: Track: Cluster: Room: Chair: Laszlo Ladanyi Chair Address: IBM Research, TJ Watson Research Ctr., PO Box 218, Rte. 134, Yorktown Heights, NY 10598 Chair E-mail: ladanyi@us.ibm.com Chair: Chair Address: Chair E-mail: SC46.1 Installation & Use of Common Optimization INterface (COIN): Windows • J. P. Fasano; IBM Research, TJ Watson Research Ctr., Rte. 134, Yorktown Heights, NY 10598; jpfasano@us.ibm.com • Laszlo Ladanyi; IBM Research, TJ Watson Research Ctr., PO Box 218, Rte. 134, Yorktown Heights, NY 10598; ladanyi@us.ibm.com • Robin Lougee-Heimer; IBM TJ Watson Research Center, PO Box 218, Rte. 134, Yorktown Heights, NY 10598-0218; robinlh@us.ibm.com • Matthew J. Saltzman; Clemson University, Dept. of Math. Sci., Clemson, SC 29634-0975; mjs@math.clemson.edu COIN is an initiative to promote open source software for OR. Bring your Windows laptop and network card to participate in this hands-on workshop using open source components in the COIN repository. No Windows-laptop? Come observe and learn more about COIN. Visit http://oss.software.ibm.com/developerworks/opensource/coin for details and prerequisites. # Marketing & Decision Making Session: SD01 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Decision Analysis Society Track: Cluster: Room: Chair: Jianmin Jia Chair Address: Chinese University of Hong Kong, Dept. of Mktg., Shatin, NT, , Hong Kong Chair E-mail: jjia@cuhk.edu.hk Chair: Chair Address: Chair E-mail: SD01.1 Retail Loyalty Programs • David Bell; Harvard Business School, Morgan Hall 171, Soldiers Field, Boston, MA 02163; dbell@hbs.edu The latest retail fashion: e-commerce, has eclipsed the last one: loyalty programs. Are they working? How are they supposed to work? The first half of the presentation will be ruminations on the issue, the second half will be results from a simple model of the situation. SD01.2 Learning-Based Theories of Reputation-Building in Repeated Games: Scaring Competitors & Building Trust • Colin Camerer; Caltech, Div. Humanities & Social Sci., Pasadena, CA 91125; camerer@hss.caltech.edu • Teck-Hua Ho; University of Pennsylvania, The Wharton Sch., Mktg. Dept., 1400 SH-DH, Philadelphia, PA 19104; hoteck@wharton.upenn.edu • Juin Kuan Chong; National University of Singapore; We extend adaptive theories of learning in repeated one-shot games to games in which one 'teacher' plays with a series of learner students, so the teacher has an economic incentive to sacrifice short-run payoffs to teach what strategies do not pay. We use 2 examples: building a reputation for toughness by fighting entry attempts and building a reputation for trust by repaying loans. SD01.3 Consumer Regret Following Switch vs. Repeat Deicsions in Outcome Sequences • J. Jeffrey Inman; University of Pittsburgh, Katz. Grad. Sch. of Bus., Pittsburgh, PA 15260; jinman@bus.wisc.edu • Marcel Zeelenberg; Tilburg University; Decision-making literature has consistently reported that decisions to maintain the status quo tend to be regretted less than decisions to switch. We examine the consequences of repeat purchases (maintaining the status quo) vs. switching in the context of information about prior consumption experiences, arguing that there are also situations where regret may be greater in the case of repeat purchases. SD01.4 Consumer Preference Uncertainty: Measures of Attribute Conflict & Extremity • Jianmin Jia; Chinese University of Hong Kong, Dept. of Mktg., Shatin, NT, , Hong Kong; jjia@cuhk.edu.hk • Mary Frances Luce; University of Pennsylvania, The Wharton Sch., Philadelphia, PA 19104; lucem@wharton.upenn.edu • Gregory W. Fischer; Duke University, Fuqua Sch. of Bus., Durham, NC 27708; fischer@mail.duke.edu We investigate preference uncertainty as a function of stimulus characteristics such as attribute conflict (discrepancy among the attributes of an alternative) and attribute extremity (very high or low attribute values). Based on a random additive multi-attribute utility model, we derive formal measures of attribute conflict and attribute extremity and test our measures empirically using consumer purchase contexts... # Computation & Graphic Layout Session: SD02 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Decision Analysis Society Track: Cluster: Room: Chair: Ross D. Shachter Chair Address: Stanford University, Dept. of MS & Eng., Serra House, Stern Hall, Stanford, CA 94305-4026 Chair E-mail: shachter@stanford.edu Chair: Chair Address: Chair E-mail: SD02.1 A Qualitative Linear Utility Theory for Spohn's Theory of Epistemic Beliefs • Phan Giang; University of Kansas, Sch. of Bus., Summerfield Hall, Lawrence, KS 66045-2003; pgiang@ukans.edu • Prakash P. Shenoy; University of Kansas, School of Bus., Summerfield Hall, Lawrence, KS 66045-2003; pshenoy@ukans.edu We formulate a qualitative 'linear' utility theory for lotteries in which uncertainty is expressed qualitatively using a Spohnian disbelief function. We argue that a rational decision maker facing an uncertain decision problem in which the uncertainty is expressed qualitatively should behave so as to maximize 'qualitative expected utility...' SD02.2 Micro-Dynamic Simulataion for Information Gathering One of the critical parts of the decision analysis is information gathering. In the world of new products and services development, there is information even beyond the expert judgement/opinion, or the situations where the expert opinions cannot be easily substantiated. We discuss micro-dynamic simulation for information gathering for decision making and the evaluation/quantification of expert opinions... SD02.3 An Influence Diagram Model for Breast Cancer Screening • Ross D. Shachter; Stanford University, Dept. of MS & Eng., Serra House, Stern Hall, Stanford, CA 94305-4026; shachter@stanford.edu • Elizabeth Burnside; Stanford University, Stanford Medical Informatics; • Daniel Rubin; Stanford University, Stanford Medical Informatics; Mammography is the best diagnostic technology currently available to decrease the mortality and morbidity of breast cancer, but mammographic interpretations and subsequent decisions vary widely among radiologists. We are developing influence diagram tools based on the standardized BI-RADS lexicon for mammography reports to improve medical decision-making. # Pricing, Incentives & Channel Policies Session: SD03 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: MSOM Track: Cluster: Room: Chair: M. Eric Johnson Chair Address: Dartmouth College, Tuck Sch. of Bus., Hanover, NH 03768 Chair E-mail: m.eric.johnson@dartmouth.edu Chair: Chair Address: Chair E-mail: SD03.1 Channel Rebates, Promotional Effort & Coordination A channel rebate is a payment from a manufacturer to a retailer based on retailer sales to consumers. We explore prospects for two commonly used forms of channel rebates, linear rebates and target rebates, to achieve channel coordination both when promotional effort influences stochastic demand and when it does not. SD03.2 Supply Chain Structures on the Internet: Marketing-Operations Coordination The development of the Internet resulted in the possibility of disintermediation of information flow and physical goods flow. As a result, alternative supply chain structures arise where the retailer is primarily concerned with the customer acquisition and the wholesaler takes care of inventory and fulfillment. We introduce and analyze such supply chains. SD03.3 Rolling Mix Strategies Rolling mix schemes where retailers receive an assortment of ever-changing products is a popular marketing strategy for toys and collectibles. We examine this strategy from a supply chain perspective and consider its applicability to other products. # MSOM Fellows Session: SD04 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: MSOM Track: Cluster: Room: Chair: Ronald G. Askin Chair Address: University of Arizona, Dept. of SIE, PO Box 210020, Tucson, AZ 85721-0020 Chair E-mail: ron@sie.arizona.edu Chair: Chair Address: Chair E-mail: SD04.1 Designing & Delivering a University Course: A Process Management Perspective • Edward A. Silver; University of Calgary, Fac. of Mgmt., 2500 University Dr. NW, Calgary, Alberta, T2N 1N4 , Canada; silver@mgmt.ucalgary.ca Based on his extensive teaching experience, the speaker believes that it can be very useful for an instructor to carefully consider the processes that are used to design and deliver a university course. This presentation will emphasize a process management perspective of the design and delivery... SD04.2 Project Work & Operations Management: Challenges & Implications • Stephen C. Graves; MIT, 77 Massachusetts Ave., Rm. E40-439, Cambridge, MA 02139-4307; sgraves@mit.edu Much of my teaching and research over the last decade involves project work. This presentation will highlight some observations and learning from these experiences. I will also discuss some implications for both how we conduct model-based research in operations management and how we teach operations management. SD04.3 Variability & Uncertainty in Manufacturing & Service Systems • John A. Buzacott; York University, Schulich Sch. of Bus., 4700 Keele St., Toronto, Ontario, M3J 1P3 , Canada; jbuzacott@ssb.yorku.ca The evolution of understanding the impact of variability and uncertainty on the performance of manufacturing and service systems along with how to design and operate them to minimize the impact will be discussed. The presentation will conclude with some conjectures about the future evolution of systems and modeling approaches. # Computational Finance Session: SD05 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Applied Probability Society Track: Cluster: Room: Chair: Benjamin Van Roy Chair Address: Stanford University, Terman 427, Stanford, CA 94307 Chair E-mail: bvr@stanford.edu Chair: Chair Address: Chair E-mail: SD05.1 Calibration in Interest Rate Markets • Yaser S. Abu-Mostafa; Caltech., 136-93, Pasadena, CA 92115; yaser@caltech.edu • Malik Magdon-Ismail; RPI, 207 Lally Bldg., CS Dept., 110 8th St., Troy, NY 12189; magdon@cs.rpi.edu We discuss the problem of calibrating multifactor Vasicek models to interest rate data; in particular, to swaps data. Of particular importance is the ability to extract the correct volatility term structure for the purposes of pricing interest rate options. We will demonstrate how conventionally used techniques are not successful, especially as the number of factors increases beyond two... SD05.2 Tax-Aware Multi-Period Portfolio Optimization • Dimitris Bertsimas; MIT, Sloan Sch. of Mgmt., E53-359, OR Ctr., 50 Memorial Dr., Cambridge, MA 02139; dbertsim@mit.edu • Georgia Mourtzinou; Dynamic Ideas, LLC, PO Box 425547, Cambridge, MA 02142; gina@aris.mit.edu We formulate the problem of optimally trading a taxable portfolio as a stochastic dynamic programming problem. Because of its dimensionality, we develop an approximation algorithm that can solve large scale problems involving thousands of assets over several periods. We show that optimal multi-period strategies outperform single period and buy-and-hold strategies. SD05.3 Pricing American Options: A Comparison of Monte Carlo Simulation Approaches • Michael C. Fu; University of Maryland, Smith School of Bus., Van Munching Hall, College Park, MD 20742-1815; mfu@rhsmith.umd.edu • Scott B. Laprise; University of Maryland, Dept. of Math., College Park, MD 20742; sbl@math.umd.edu • Dilip B. Madan; University of Maryland, Smith Sch. of Bus., College Park, MD 20742; dmadan@rhsmith.umd.edu • Yi Su; University of Maryland, Smith Sch. of Bus., College Park, MD 20742; ysu@glue.umd.edu • Rongwen Wu; University of Maryland, Dept. of Math., College Park, MD 20742; rxw@math.umd.edu We compare various Monte Carlo simulation-based approaches for pricing American-style derivatives on a common set of numerical problems. We also introduce another simulation-based approach that employs a simultaneous perturbation stochastic approximation algorithm. SD05.4 Regression Methods for Pricing Complex American-Style Options • Benjamin Van Roy; Stanford University, Terman 427, Stanford, CA 94307; bvr@stanford.edu A number of researchers have proposed methods for approximating pricing functions of high-dimensional American-style options. We discuss characteristics common to these methods and a key ingredient possessed by some that reduces approximation error dramatically. We also propose extensions to the methodology that should lead to further computational advantages. # Travel Models for Planning & Evaluation Session: SD06 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Transportation Science Section Track: Cluster: Room: Chair: Chandra Bhat Chair Address: University of Texas, Dept. of Civil Eng., ECJ 6.810, Austin, TX 78712 Chair E-mail: bhat@mail.utexas.edu Chair: Chair Address: Chair E-mail: SD06.1 Destination Choice Modeling for Home-Based Recreation Trips: Analysis & Implications for Land-Use, Transportation & Air Quality Planning Attraction-end choice for home-based urban recreational trips is estimated using a non-linear-in-parameters multi-nomial logit model. The effects of level-of-service, zonal attributes, trip attributes and socio-demographics on recreational attraction-end choice are examined and the implications on land-use and transportation planning and air quality analysis are discussed. SD06.2 A Hierarchical Approach for Performance & Institutional Evaluation • Stephen P. Mattingly; University of Alaska, Dept. of Civil & Environ. Eng., PO Box 755900, Fairbanks, AK 99775-5900; ffspm@aurora.uaf.edu • R. Jayakrishnan; University of California, Dept. of Civil & Environ. Eng., Irvine, CA 92697; • Michael G. McNally; University of California, Dept. of Civil & Env. Eng., Irvine, CA 92697; We offer a new hierarchical approach for transportation evaluation. Decision-theory can be adapted through this approach to properly assess qualitative and institutional issues as well as technical factors. The technique simplifies the process and creates an overall framework that allows for comparisons typically not considered in cost-benefit analysis. SD06.3 Emerging Dynamic Traffic Assignment Methods for Planning & Evaluating Transportation Network Operational Strategies • Hani S. Mahmassani; University of Texas, Dept. of Civil Eng., ECJ 6.2, Austin, TX 78712-1076; masmah@mail.utexas.edu • Yi-Chang Chiu; University of Texas, Dept. of Civil Eng., ECJ 6.2, Austin, TX 78712; • Ahmded F. Abdelghany; University of Texas, Dept. of Civ. Eng., ECJ 6.2, Austin, TX 78712; afaissal@mail.texas.edu • Khaled F. Abdelghany; University of Texas, Dept. of Civil Eng., ECJ 6.2, Austin, TX 78712; kfaissal@mail.utexas.edu • Nhan Huynh; University of Texas, Dept. of Civ. Eng., ECJ 6.2, Austin, TX 78712; • Xuesong Zhou; University of Texas, Dept. of Civ. Eng., ECJ 6.2, Austin, TX 78712; We present an overview of DYNASMART-P, a new generation of tools, based on dynamic traffic assignment methodologies to support transportation network planning and operations decisions. Different aspects of DYNASMART-P including methodological issues, modeling framework and model features and capabilities are presented. SD06.4 A Freight Distribution Model for County-to-County Commodity Flows • Aruna Sivakumar; University of Texas, 109 West 39th St., Apt. 111, Austin, TX 78751; arunas@mail.utexas.edu • Chandra Bhat; University of Texas, Dept. of Civil Eng., ECJ 6.810, Austin, TX 78712; bhat@mail.utexas.edu A quasi-likelihood approach is formulated for modeling the fraction of freight produced at an origin node that is destined to each of several consumption nodes. The approach is applied to analyze county-to-county freight commodity flows within Texas and between Texas counties and neighboring states/Mexico. The Reebie freight database is used in the analysis. # Improving Intuition through Interactive Models Session: SD07 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: INFORM-ED Track: Cluster: Room: Chair: Sam L. Savage Chair Address: Stanford University, Dept. of MSE, 417 Terman Eng. Ctr., Stanford, CA 94305-4023 Chair E-mail: savage@stanford.edu Chair: Chair Address: Chair E-mail: SD07.1 OR in Pictures • Jeffrey D. Camm; University of Cincinnati, QAOM Dept., ML 0103, Cincinnati, OH 45221-0130; jeff.camm@uc.edu Through a portfolio of examples, we illustrate how to easily use the visualization capabilities in Microsoft Excel to enhance student understanding of basic MS/OR concepts. SD07.2 Integrated Excel Text & Demonstrations • David W. Ashley; University of Missouri, Bloch Sch. of Bus./Pub. Admin., 5100 Rockhill Rd., Kansas City, MO 64110-2499; ashleyd@umkc.edu Text, text boxes, comments, visual basic, formulas and graphics can all be used within Excel to create pedagogical documents that enhance student learning. We address instructional materials dealing with queueing, simulation, optimization and mathematics review, as well as pros and cons relative to other formats. SD07.3 Blitzograms: Interactive Histograms • Sam L. Savage; Stanford University, Dept. of MSE, 417 Terman Eng. Ctr., Stanford, CA 94305-4023; savage@stanford.edu By interacting with probability distributions, can we learn more about them? # Business Applications Session: SD08 Date/Time: Sunday 15:00-16:30 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Hernan Wurgaft Chair Address: Industria Technologies LLC, 5116 Simpson Lake Rd., West Bloomfield, MI 48323 Chair E-mail: hernanw@industriatechnologies.com,, http://www.industriatechnologies.com Chair: Chair Address: Chair E-mail: SD08.1 Cadet Summer Research Program: A Lesson in how Forecasting is Really Done The USAFA provides Cadets the opportunity to apply OR techniques with both military and non-military organizations. My recent experience in business shows the random guessing game that one corporation uses to predict future sales and the strides they are taking to incorporate formal forecasting and modeling techniques into their decision making. SD08.2 A Decision-Analytic Framework for Evaluating Natural Gas Storage Value Strategic valuation of natural gas storage presents decision-makers charged with managing such assets with a host of challenges. In systematically managing gas storage, decision-makers must consider current market prices, as well as the possibility of favorable price movements during the valuation period. The stochastic nature of gas prices, however, makes reliable inferences about market dynamics inherently difficult... SD08.3 Determining Optimal Daily Staffing Levels for the Whistler/Blackcomb Ski School • Martin L. Puterman; University of British Columbia, Ctr. for Op. Excellence, 2053 Main Hall, Vancouver, BC, V6T 1Z2 , Canada; marty@coe.ubc.ca • David W. Glenn; University of British Columbia, Ctr. for Op. Excellence, 2053 Main Mall, Vancouver, BC, V6T 1Z2 , Canada; dglenn@coe.ubc.ca • Stanley Tse; University of British Columbia, Ctr. for Op. Excellence, 2053 Main Mall, Vancouver, BC, V6T 1Z2 , Canada; stse@coe.ubc.ca • Isabelle Smith; University of British Columbia, Ctr. for Op. Excellence, 2053 Main Mall, Vancouver, BC, V6T 1Z2 , Canada; ismith@coe.ubc.ca The daily variability of lesson volume makes it difficult and time consuming for management to determine the number of instructors to schedule. We describe models developed to set daily staffing levels based on demand forecasting and newsvendor inventory models and a computer-based tool developed for use in the 2000/2001 season. SD08.4 Formulating Corporate Strategy with Collaborative Scenarios A common problem in using mathematical models to formulate corporate strategy is how to supply the models with meaningful information consistently. This information is contributed by executives in different functional and geographical areas. We discuss software architecure and business processes that are required to collectively build scenarios and formulate strategy. # Applications of Revenue Management II Session: SD09 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Revenue Management Section Track: Cluster: Room: Chair: Ying Kang Chair Address: PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006 Chair E-mail: ykang@prosrm.com Chair: Chair Address: Chair E-mail: SD09.1 Revenue Management for Tourism Small- & Medium-Sized Enterprises • Michael Mulvey; Dublin Institute of Technology, Dublin 1, , Ireland; http://www.dit.ie The tourism industry in the European Union is dominated by SMEs that make up 95% of the sector. RM is one area where SMEs are endeavoring to deploy in a way which is meaningful and feasible for small-scale operators. We propose a model of how RM can be used in the context of e-commerce and tourism destination management systems. SD09.2 Extensions of Dynamic Pricing in Manufacturing & Retail Industries Dynamic pricing has become a popular tool in manufacturing and retail industries to increase profits and decrease demand variability. We extend dynamic pricing models to include considerations such as lead time and multiple products and present new computational analysis regarding the impact of dynamic pricing on the supply chain. SD09.3 Price-Directed Coordination of Replenishments The idea of price-directed control is to use a policy that exploits optimal dual prices from a mathematical programming relaxation of the underlying control problem. In the context of industrial gas distribution, we use prices to coordinate inventory replenishments over time and across items/locations so as to minimize replenishment costs. SD09.4 PROS Capacity Management System: A Real Revenue Management Application & Implementation • Ying Kang; PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006; ykang@prosrm.com • Viroj Buraparate; PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006; vburaparate@prosrm.com We will describe and introduce the first successful application and implementation of RM system for energy industry by PROS Revenue Management Inc. Some interesting discussions and findings of modeling research and business development efforts are also included. # Optimization & Scheduling II Session: SD10 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Scheduling Room: Chair: Qing Liu Chair Address: AT&T Bell Laboratories, 379 Campus Dr., Rm. 2B212, Somerset, NJ 08873 Chair E-mail: qliu@att.com Chair: Chair Address: Chair E-mail: SD10.1 A Partnership Selection Model for Supply Chains We present a mathematical model for strategic selection of both upstream and downstream partners for supply chains with an objective to improve the supply chain performance. The model can be used to estimate the tradeoffs between overall cost savings along the supply chain and the lead-time from raw materials to end product to customers. SD10.2 An Effective Algorithm for the Robotic Cell Cyclic Scheduling Problem with Two Machines & Time Window Constraints • Qing Liu; AT&T Bell Laboratories, 379 Campus Dr., Rm. 2B212, Somerset, NJ 08873; qliu@att.com We propose a strong polynomial algorithm to solve the robotic cell cyclic scheduling problem with 2 machines and time window constraints. A 2-step procedure is presented to search for the optimal solution. An extension of the algorithm to solve the general robotic cyclic scheduling problem is discussed. SD10.3 Scheduling the Workforce to Maximize Employee Satisfaction • Anna Olecka; Rutgers University, RUTCOR, 640 Bartholomew Rd., New Brunswick, NJ 08901; Improved retention is an important business goal in an environment with part -time workforce. High turnover increases training cost and decreases productivity. We present a 2-phase approach to a problem of optimal workforce scheduling. Phase 1 determines shift size, phase 2 assigns workers to shifts while optimizing their preferences. SD10.4 Scheduling in Lean Manufacturing Environments • Shrikant S. Panwalkar; Purdue University, Krannert Sch. of Mgmt., 1310 Krannert Bldg., West Lafayette, IN 47907-1310; panwalkars@mgmt.purdue.edu Lean manufacturing makes use of pull systems as opposed to push systems used in conventional manufacturing. An essential component of a lean manufacturing system is JIT; this in turn makes transportation and logistics different from the conventional system. Is scheduling done differently in lean manufacturing? We attempt to look at this question and find new areas for scheduling research. # Combinatorial (Bundle) Auctions Session: SD11 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Integer Programming Room: Chair: Steef L. van de Velde Chair Address: Erasmus University, Rotterdam Sch. of Mgmt., PO Box 1738, Rotterdam, 3000 DR , The Netherlands Chair E-mail: svelde@fac.fbk.eur.nl Chair: Chair Address: Chair E-mail: SD11.1 Combinatorial Auctions from a Primal-Dual Perspective • Andreas S. Schulz; MIT, Sloan Sch. of Mgmt. & OR, BLdg. E53-361, 30 Acorn St., Cambridge, MA 02142-1320; schulz@mit.edu • Rudolf Muller; University of Maastricht, Dept. of Quantitative Econ., Maastricht, 6200 MD , The Netherlands; r.muller@ke.unimaas.nl An important aspect of the design of combinatorial auctions is the winner determination problem. We take a look at primal-dual algorithms, which share 3 favorable properties: computed solutions are supported by individual item prices, corresponding payment schemes may enforce truth revelation and certificates of optimality of assignments are immediately available. SD11.2 Complexity & Algorithms for Winner Assignment in Combinatorial Auctions • Rudolf Muller; University of Maastricht, Dept. of Quantitative Econ., Maastricht, 6200 MD , The Netherlands; r.muller@ke.unimaas.nl • Stan van Hoesel; University of Maastricht, Dept. of Quantitative Econ., Maastricht, 6200 MD , The Netherlands; s.vanhoesel@ke.unimaas.nl We present an analysis of the complexity of the problem to assign bids to bidders in combinatorial auctions. We show that the case of identical assets can be solved in polynomial time. We give some computational results using integer linear programming formulations and heuristics SD11.3 The Winners Determination Problem in Tendering Transportation Services • Steef L. van de Velde; Erasmus University, Rotterdam Sch. of Mgmt., PO Box 1738, Rotterdam, 3000 DR , The Netherlands; svelde@fac.fbk.eur.nl • Roelof Kuik; Erasmus University, Rotterdam Sch. of Mgmt., PO Box 1738, Rotterdam, 3000 DR , The Netherlands; rkuik@fac.fbk.eur.nl • Linda van Norden; Erasmus University, Rotterdam Sch. of Mgmt., PO Box 1738, Rotterdam, 3000 DR , The Netherlands; lnorden@fac.fbk.eur.nl The tendering process for outsourcing transportation of bulk chemicals can be seen as a combinatorial auction. We present an algorithm for the solution of the winners determination problem. # SCM: Innovative Solutions, Global Decisions Session: SD12 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Logistics & Supply Chain Management Room: Chair: Lisa A. Ferguson Chair Address: Hofstra University, Dept. of Mgmt. & General Bus., 228 Weller Hall, Hempstead, NY 11549 Chair E-mail: Chair: Chair Address: Chair E-mail: SD12.1 A Model for Supply Chain Performance Measurement • Abby Ghobadian; Middlesex University, The Business Sch., The Burroughs, London, NW4 4BT , UK; a.ghobadian@mdx.ac.uk • David Gallear; Middlesex University, The Business Sch., The Burroughs, London, NW4 4BT , UK; d.gallear@mdx.ac.uk • Rose Li; Middlesex University, The Business Sch., The Burroughs, London, NW4 4BT , UK; r.li@mdx.ac.uk Increased global competition and the perceived advantage in purchasing enjoyed by Japanese companies have heightened focus on SCM. A review of the literature suggests that there is no agreement as to the boundary of SCM. We have used Thomas & Griffin's (1996) stages of SCM: procurement, production and distribution. Each stage normally consists of several sub-activities... SD12.2 Demand Side Performance Measures for the Supply Chain • Savas Ozatalay; Widener University, 1 University Place, Sch. of Bus. Admin., Chester, PA 19013-5792; Most of the research in supply chain is contained in the supply side. Many companies that excel in producing their products experience financial difficulties. Many of these difficulties come from the mismanagement of the demand side. This study develops a series of performance measurements that companies can use in synchronizing demand and supply sides of the supply chain. SD12.3 Breakthrough Business Models & their Creative Supply Chain Solutions The emergence of new operating business models specifically direct to business/consumer distribution channels, mass customization, and high velocity supply chains, call for innovative process and internet supply chain solutions. During this presentation the speaker will provide both the challenges of managing in a direct environment along with leading edge supply chain management practices to ensure a positive customer experience while achieving balanced financial results. SD12.4 Structuring Global Supply Chain Management Decisions • Lisa A. Ferguson; Hofstra University, Dept. of Mgmt. & General Bus., 228 Weller Hall, Hempstead, NY 11549; This research provides guidelines for making global SCM decisions. The decisions are divided into the categories of strategic, tactical and operational decisions. We also present a comprehensive list of factors to consider when making global SCM decisions. # Recent Advances in Manufacturing Scheduling Session: SD13 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Manufacturing & Logistics Room: Chair: Philippe Baptiste Chair Address: University of Technology of Compiegne, HeuDiaSyC, Ctr. Recherches de Royallieu, Compiegne Cedex, 60205 , France Chair E-mail: philippe.baptiste@hds.utc.fr Chair: Chair Address: Chair E-mail: SD13.1 Scheduling Identical Jobs on a Multi-Purpose Machine with Uniform Speed • Emmanuel Neron; Universite de Tours, Laboratoire d'Info., 64 av. Jean Portalis, Tours, F-37200 , France; • Alain Eloundou; Universite de Tours, Laboratoire d'Info., 64 av. Jean Portalis, Tours, F-37200 , France; The problem we study deals with a real-life farming problem. For each parcels of a field, 3 operations have to be processed. These operations are submitted to precedence constraints and 2 types of machines are available. Our goal is to minimize the makespan. Heuristics and exact models will be presented. SD13.2 Scheduling & Work-in-Process Control: An On-Line Approach • Fabrice Chauvet; Bouygues Telecom, Europa L, 30 av. de l'Europe, Velizy, 78944 , France; We propose to analyze a no-wait controllable processing times production systems. In such a system, the processing times can be selected in given time intervals. The proposed methodology allows one to control the WIP and production cycle time. The system productivity is optimized through original on-line algorithms. SD13.3 Slack-Based Techniques for Robust Scheduling • Andrew J. Davenport; IBM TJ Watson Research Center, PO Box 218, Rte. 134, Yorktown Heights, NY 10598; We present a number of techniques for generating robust schedules in the presence of stochastic machine breakdowns. These techniques make use of statistical characterizations of breakdowns to determine how much slack time to incorporate into a schedule. We present results of a simulation study of these techniques. SD13.4 Exploiting Activity Interactions in Constraint-Based Scheduling • Philippe Laborie; ILOG, 9 rue de Verdun, Gentilly Cedex, , France; The properties that are sufficient to characterize a solution to a finite capacity scheduling problem fall into 2 categories: the ones based on the absolute positions of activities along the time axis (resource usage profile) and the ones that rely on the interactions between activities competing for the same resource (precedences between activities start/end times). We discuss, classify and compare theconstraint programming techniques built on this second category of properties. # Airline Schedule Planning Session: SD14 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Aviation Applications Section/Transportation Science Section Track: Cluster: Room: Chair: Diego Klabjan Chair Address: University of Illinois, Dept. of MIE, 1206 West Green St., Urbana, IL 61801 Chair E-mail: klabjan@uiuc.edu Chair: Chair Address: Chair E-mail: SD14.1 Robust Crew Scheduling: Move-Up Crews Due to the flight disruptions in operations, the crew scheduling cost at the end of a month is substantially higher than the projected cost in planning. We present a model and solution methodologies that produce more robust crew schedules in planning. We propose an objective function that captures the number of move-up crews, i.e. the crews that can be swapped in operations... SD14.2 Scheduling Strategies for Improving Operational Robustness • Yanna Ageeva; MIT International Center for Air Transportation, 77 Massachusetts Ave., Cambridge, MA 02139; yana99@mit.edu • John-Paul Clarke; MIT International Center for Air Transportation, Aeronautics/Astronautics Dept., 77 Massachusetts Ave., Cambridge, MA 02139-4307; johnpaul@mit.edu Lack of robust real-time decision-making tools in airline operations control centers has resulted in poor operational performance when faced with operational delays from severe weather and unexpected aircraft and personnel failures. Presently, airlines create 'optimal' schedules that don't take into consideration possible weather and air traffic control delays, except through the addition of increased time buffers in block times and turn times... SD14.3 Airline Fleet Assignment: An Alternative Model & Solution Approach • Amr Farahat; MIT, OR Ctr., E40-130, Cambridge, MA 02139; afarahat@mit.edu • Cynthia Barnhart; MIT, OR Ctr., Rm. 1-229, 77 Massachusetts Ave., Cambridge, MA 02139; cbarnhar@mit.edu • Manoj Lohatepanont; MIT, Ctr. for Transport. Studies, Dept. of Civil & Environ. Eng., Cambridge, MA 02139; lmanoj@mit.edu We consider the airline fleet assignment problem where the objective is to determine a profit-maximizing assignment of fleet types to flight legs. We present and analyze a new formulation that more accurately models passenger spill and recapture, without some of the common associated computational difficulties. We discuss our new approach and provide preliminary results using data from a major airline. SD14.4 Parc AirPlanner: Revenue ReTimer Module Revenue ReTimer, the next generation of optimisers, focuses on refining flight timings with a view to maximising revenue. It is capable of handling slot holdings and risk, fleet availability and the main schedule feasibility issues. It has passed the first development milestone in June 2000 and testing on real data, involving up to 2500 flights in a standard week view, has shown revenue improvements of up to 20% using the tool. # Collaboration & Competition in Supply Chain Management Session: SD15 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Quantitative Models in Supply Chain Management Room: Chair: Greys Sosic Chair Address: University of British Columbia, Fac. of Commerce & Bus. Admin., Operations & Logistics Div., Vancouver, BC, V6T 1Z2 , Canada Chair E-mail: greys.sosic@commerce.ubc.ca Chair: Chair Address: Chair E-mail: SD15.1 Coordination Buyback & Competition • Mahesh Nagarajan; University of Southern California, Dept. of Info. & Op. Mgmt., Marshall Sch. of Bus., Los Angeles, CA 90089; • Yehuda Bassok; University of Southern California, Dept. of Info. & Op. Mgmt., Marshall Sch. of Bus., Los Angeles, CA 90089; We study 'buyback' policies as a tool for channel coordination. We study a system with 1 manufacturer and 1 retailer. Using the Nash-bargaining concept, we show that buybacks are likely to be used when the retailer is 'weak'. We then study a system with 2 competing manufacturers. We show that when competition is intense at equilibrium, buybacks are not likely to be used. SD15.2 A Three-Stage Model of a Decentralized Distribution System of Retailers • Greys Sosic; University of British Columbia, Fac. of Commerce & Bus. Admin., Operations & Logistics Div., Vancouver, BC, V6T 1Z2 , Canada; greys.sosic@commerce.ubc.ca • Daniel Granot; University of British Columbia, Fac. of Commerce & Bus. Admin., Operations & Logistics Div., Vancouver, BC, V6T 1Z2 , Canada; We study the challenges involved in achieving both collaboration and truthful revelation of information about residual inventories in a 3-stage model of a decentralized distribution system. The distribution system consists of n retailers, each of whom faces a stochastic demand for an identical product and whom are willing to share residual inventories in order to increase profit. SD15.3 The Value of Information Sharing in a Two-Stage Supply Chain with Production Capacity Constraints • David Simchi-Levi; MIT, Dept. of Civil & Environ. Eng., 77 Massachusetts Ave. Rm. 1171, Cambridge, MA 02139; dslevi@mit.edu • Yao Zhao; Northwestern University, Evanston, IL 60208; yzh841@casbah.acns.nwu.edu We consider a simple 2-stage supply chain with a single retailer facing iid demand and a single manufacturer with finite production capacity. We study the impact of information sharing on cost and service level as a function of the production capacity, and the frequency and timing in which demand information is transferred to the manufacturer. SD15.4 Price & Delivery Frequency Competition in a Supply Chain • Albert Y. Ha; Yale University; • Lode Li; Yale University; • Shu-Ming Ng; Hong Kong University of Science & Technology; We consider a system in which 2 suppliers compete for supplies to a manufacturer and analyze horizontal and vertical competition under different assumptions on how pricing and delivery frequency decisions are made. Our analysis sheds light on the strategic role of delivery frequency and the practice of JIT delivery. # Topics in Multicriteria Decision Making Session: SD16 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: MCDM Room: Chair: Anne Davey Chair Address: Northeastern State University, Coll. of Bus. & Industry, Tahlequah, OK 74464 Chair E-mail: davey@cherokee.nsuok.edu Chair: Chair Address: Chair E-mail: SD16.1 The Elimination of Tradeoffs in Modern Business & Economics Tradeoff analysis is becoming one of the most important parts of modern business and economic decision making. We describe the global network economy and its need for tradeoff elimination on the part of producers and service providers. We then offer some conceptual and analytical tools of De novo programming for dealing with this goal. SD16.2 Does the Choice of a Multiple Criteria Method Really Matter? • Adel Guitouni; Defense Research Establishment Valcartier, Decision Support Tech. Section, 2459 Pie-XI Nord, Val-Belair, Quebec, G3J 1X5 , Canada; adel.guitouni@drev.dnd.ca Multi-criteria decision aid literature includes a panoply of discrete methods. Does the choice of a method for a specific decision really matter? We propose a framework based on MCDA methods and decision making situations. Since MCDA methods differ primarily at the level of aggregation, we consider multi-criteria aggregation procedures (MCAP). SD16.3 Preference Issues in Group Decision Making • David L. Olson; Texas A&M University, IOM Dept., 4217 TAMUS, College Station, TX 77843-4217; dolson@tamu.edu SD16.4 Do Multiple Criteria Need to be Independent? • Larry Jenkins; Royal Military College of Canada, Dept. of Bus. Admin., PO Box 17000 Stn Forces, Kingston, Ontario, K7K 7B4 , Canada; jenkins-l@rmc.ca In multi-criteria decision analysis, one aims to define the criteria to be as independent as possible. However, often the criteria that are most obvious, definable and measurable are not independent. We use a number of examples to explore where this lack of independence may be a problem. # Network Flows: Freight Applications Session: SD17 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Optimization Section Track: Cluster: Network Flows Room: Chair: Tassio Carvalho Chair Address: IBM TJ Watson Research Center, PO Box 218, Yorktown Heights, NY 10598 Chair E-mail: tassio@us.ibm.com Chair: Chair Address: Chair E-mail: SD17.1 An Inverse Optimization Algorithm for Calibrating Railroad Block Impedances • Jeff D. Day; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205; • George L. Nemhauser; Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205; george.nemhauser@isye.gatech.edu • Joel S. Sokol; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205; jsokol@isye.gatech.edu To calculate railroad block impedances so that each commodity routes on its unique shortest path, we consider a variation of the inverse multicommodity flow problem. We approximate a blocking plan with several aggregated networks. Because a given block may be shared between multiple networks, the joint optimization SD17.2 Composite Variable Formulations for Express Shipment Service Network Design • Cynthia Barnhart; MIT, OR Ctr., Rm. 1-229, 77 Massachusetts Ave., Cambridge, MA 02139; cbarnhar@mit.edu • Andrew P. Armacost; MIT, OR Ctr., Rm. E40-139, 77 Massachusetts Ave., Cambridge, MA 02139; armacost@mit.edu We consider the problem of simultaneously determining aircraft routes, fleetings and packages flows for an overnight package carrier. We derive a composite variable formulation that implicitly captures package flows, detail the relationship between this strategy and traditional network design approaches and present computational results for real-world problem instances. SD17.3 A Decomposition Algorithm for the Deterministic Dynamic Multicommodity Allocation of Empty Containers • Jawad Abrache; DIRO/Universite de Montreal, CRT, CP 7128 Succ Centre-Ville, Montreal, Quebec, H3C 3J7 , Canada; • Teodor Gabriel Crainic; DIRO/Universite de Montreal, CRT, CP 7128 Succ Centre-Ville, Montreal, Quebec, H3C 3J7 , Canada; theo@crt.umontreal.ca • Michel Gendreau; DIRO/Universite de Montreal, CRT, CP 6128, succ. Centre-ville, Montreal, Quebec, H3C 3J7 , Canada; michelg@crt.umontreal.ca The allocation of empty containers to customer requests is an important problem encountered by container shipping companies in the management of their transportation operations. The problem addresses the short-term planning of empty container movements intended to satisfy customers requests and the need to reposition empties for future demand. We consider a dynamic model proposed for the deterministic formulation of the problem... SD17.4 A Strategic Planning Model for River Barge Businesses • Richard C. Larson; MIT, Ctr. for Advanced Ed. Services, Bldg. 9, Rm. 215, Cambridge, MA 02139; rclarson@mit.edu A circulation flow linear program is developed to select optimal markets for river barges. The solution that maximizes net revenue routes the fleet of barges strategically over all alternative routes and product loads and includes deadheading. Computational results are included. # The Web: Consumers, Sites & Marketing Strategies Session: SD18 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: E-Commerce Room: Chair: Patrali Chatterjee Chair Address: Rutgers University, Dept. of Mktg., 180 University Ave., Newark, NJ 07102-1897 Chair E-mail: patrali@andromeda.rutgers.edu Chair: Chair Address: Chair E-mail: SD18.1 Optimizing Web Sites using Randomized Experiments • Charles F. Hofacker; Florida State University, Coll. of Bus., Tallahassee, FL 32306-1110; chofack@cob.fsu.edu Two important features of the WWW are it's dynamic nature and that visitors or consumers leave a digital record on the server. In combination, these characteristics allow marketing scientists to create elegant randomized experiments that can yield rapid and actionable results. We summarize the methodology and some early empirical findings. SD18.2 Consumer Response to Web Site Privacy Policies • Lisa R. Klein; Rice University, Jones Grad. Sch., 307 Herring, 6100 Main St., Houston, TX 77005; • Tracy B. Dennis; Rice University, Jones Grad. Sch., 225 Herring, 6100 Main St., Houston, TX 77005; dennist@rice.edu • Tenley O'Shaughnessy; Rice University, 6100 Main St., Houston, TX 77005; • Deb Paul; Rice University, 6100 Main St., Houston, TX 77005; Most current research exploring consumer privacy concerns on the web relies upon surveys. The research presented here examines consumer behavioral response to web site privacy claims through a controlled experiment. Results suggest that consumers do not rely heavily on web site privacy claims when evaluating the trustworthiness of a web site. SD18.3 The Interactive Brand: The Role of Brand in Network Economy & Infomediation Environments • Andreina Mandelli; Universita L. Bocconi, Universita della Svizzera, Italiana, Lugano CH, Milan, , Italy; mandelli@mktgsda.inet.it Brand in commerce has a special role as a cognitive shortcut for the customer, building both awareness and trust. We argue that in the digital markets, this role is still important but brands must change the rules of relationships with customers, particularly regarding the symmetry of information and power. SD18.4 Dotcoms vs. Traditionals? Trends & Alliances in Internet Retailing Dotcom 'pure plays' such as Amazon.com sell traditional goods, while traditional retailers, e.g., Barnes & Noble, with different core competencies, embrace the Internet channel. Which will dominate the channel and which will dominate the market? Alliances - bricks & clicks, rather than bricks vs. clicks - are appearing and may rule the future. Industry and alliance data illuminate these issues. # Forecast-Facilitated Operations in Supply Chains Session: SD19 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Technology Innovations & Operations Room: Chair: Layth C. Alwan Chair Address: University of Wisconsin, School of Bus., Milwaukee, WI 53201 Chair E-mail: alwan@uwm.edu Chair: Chair Address: Chair E-mail: SD19.1 Inventory Strategies of Online Retailers • Frank Y. Chen; National University of Singapore, Dept. of Dec. Sciences, 10 Kent Ridge Crescent, Singapore, 119260 , Singapore; fbachen@nus.edu.sg • S. H. Hum; National University of Singapore, Business Sch., 17 Law Link, Singapore, 119260 , Singapore; • C. H. Sim; National University of Singapore, Business Sch., 17 Law Link, Singapore, 119260 , Singapore; We consider inventory strategies of Internet retailers. The retailer faces options of holding her own inventory or outsourcing through the third party(ies). We evaluate these inventory strategies through mathematical modeling and numerical experiments. Built upon simple OR models, the numerical experiments reveal patterns consistent with those being practiced in the industry. SD19.2 Supply Channel Design Based on Projected Market Diffusion • Dong-Qing Yao; University of Wisconsin, Milwaukee, WI 53201; dyao@uwm.edu • John J. Liu; University of Wisconsin, Sch. of Bus., Milwaukee, WI 53201; jjl@uwm.edu The manufacturer's decision to sell directly to end-users has been the most controversial issue in supply chain reengineering. Some suggest the e-channel be the only choice, while just view it as one additional choice. We characterize the projected diffusion under e-channel and analyze whether the company can benefit from keeping both options. If yes, then how. SD19.3 Forecast-Based Order Policies & Bullwhip Effect under AR(1) Demand • Layth C. Alwan; University of Wisconsin, School of Bus., Milwaukee, WI 53201; alwan@uwm.edu • John J. Liu; University of Wisconsin, Sch. of Bus., Milwaukee, WI 53201; jjl@uwm.edu • Dong-Qing Yao; University of Wisconsin, Milwaukee, WI 53201; dyao@uwm.edu We examine the relationship of forecast-based order policies and bullwhip effect in a simple supply chain under an AR(1) market demand. We show that under a direct-forecast ordering policy, the bullwhip effect can be subdued over time. We also demonstrate that the underlying inventory follows a non-stationary ARIMA process. Examples and comparisons with other common forecast-based order policies are presented. SD19.4 Statistical Quality Methods to Reduce the Bullwhip Effect An important supply chain research problem is the bullwhip effect caused by information distortion and variation amplification along a supply chain, which can lead to tremendous inefficiencies such as excessive inventory investment and lost revenues. Motivated by statistical quality methods, we propose a class of order-up-to policies and develop a nearly optimal policy to reduce the bullwhip effect. # Tabu Search in Design Problems Session: SD20 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Tabu & Scatter Search Room: Chair: Jose-Luis Gonzalez-Velarde Chair Address: Monterrey Tech., Garza Sada 2501, Monterrey NL, 64849 , Mexico Chair E-mail: lugonzal@campus.mty.itesm.mx Chair: Chair Address: Chair E-mail: SD20.1 Robust Capacitated Network Design • Ada M. Alvarez; Universidad Autonoma de Nuevo Leon, Pedro de Alba s/n, Cuidad Universitaria, San Nicolas de Garza, CP 66450 , Mexico; Designing a minimum cost network is a fundamental combinatorial optimization problem that arises in a wide variety applications;this is known to be NP-hard. The problem addresses how to configure a network and accounts for the fixed costs of arcs chosen to be in the network as well as the cost of routing goods through the network defined by the arcs... SD20.2 Optimization of a Two-Dimensional Structure with Tabu Search • Edmundo T. Zavala; Universidad de Sonora, Dept. Ing. Civil y Minas, Av. Rosales S/N Col. Centro, Hermosillo, Sonora, CP 83000 , Mexico; • Pedro Flores Perez; Universidad de Sonora, Dept. de Matematicas, Av. Rosales S/N Col. Centro, Hermosillo, Sonora, CP 83000 , Mexico; plores@gauss.nat.uson.mx • Angel Kuri Morales; Instituto Politecnico Nacional, Ctr. Investigacion/Computacion, Distrito Federal, , , Mexico; akuri@pollux.cic.ipn.mx • Jose-Luis Gonzalez-Velarde; Monterrey Tech., Garza Sada 2501, Monterrey NL, 64849 , Mexico; lugonzal@campus.mty.itesm.mx Given a collection of joints and a corresponding collection of members connecting pairs of joints, find the cross sectional area of each member so the resulting structure has minimal weight and can accommodate a given load at each joint. We consider the case where the members are selected from a catalogue; this constraint makes this problem a discrete optimization problem... SD20.3 Tabu Search for Wavelength Assignment Planning in WEM Optical Networks • Ramon Rodriguez-Dagnino; Monterrey Tech., Garza Sada 2501, Monterrey NL, 64849 , Mexico; • Jose-Luis Gonzalez-Velarde; Monterrey Tech., Garza Sada 2501, Monterrey NL, 64849 , Mexico; lugonzal@campus.mty.itesm.mx We study a TS heuristic procedure for wavelength assignment is WDM multihop optical networks. The selection of the route with minimum delay is stated as a discrete optimization problem by several authors in the literature; however, most of the studies have focused on small number of nodes and TS has been barely considered for these applications. We report several experiments. # Session IV Session: SD21 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: SICUP Room: Chair: Elita A. Mukhacheva Chair Address: Ufa State Aviation Technical University, Karl Marx Str. 12, Ufa, Bashkortostan, 450000 , Russia Chair E-mail: elita@vmk.ugatu.ac.ru Chair: Chair Address: Chair E-mail: SD21.1 Densest Translational Polygon Lattice Packing using Full Dimensional Cutting Planes An algorithm and implementation is given for densest translation lattice packing of non-convex polygons: a layout which can be replicated without overlap at each point of a planar lattice. This algorithm is the first to use full dimensional cutting planes. It has useful applications in industry, particularly clothing manufacturing. SD21.2 Optimal Double Lattice Placement of Polygons on a Plane • Yu Stoyan; Ukrainian National Academy of Science, Math Modeling & Optimal Design, Mech. Eng. Problems Inst., Kharkov, 610046 , Ukraine; stoyan@ipmach.kharkov.ua • V. N. Patsuk; Ukrainian National Academy of Science, Math Modeling & Optimal Design, Mech. Eng. Problems Inst., Kharkov, 610046 , Ukraine; Considering translates of the polygons S and M by linear combinations of basis vectors a, b, g with coefficients no more than 1 ensure non-overlapping. The density maximum belongs to a 1-dimensional set (polygons being are congruent to 2 different polygons). In a special case simultaneous touching of S and S(a), M and M(b) can describe the solution. SD21.3 Placement of Rectangles & Regular Polygons into a Strip taking into Account Errors of the Initial Data Problem • Yu Stoyan; Ukrainian National Academy of Science, Math Modeling & Optimal Design, Mech. Eng. Problems Inst., Kharkov, 610046 , Ukraine; stoyan@ipmach.kharkov.ua • T. Romanova; Ukrainian National Academy of Science, Math Modeling & Optimal Design, Mech. Eng. Problems Inst., Kharkov, 610064 , Ukraine; sherom@kharkov.ua It is necessary to place a set of objects (regular polygons or rectangles) into a strip taking into account size errors so that a length of the occupied part of the strip is minimal. On the basis of geometric design and interval analysis theory, solution methods are developed. SD21.4 no show • Elita A. Mukhacheva; Ufa State Aviation Technical University, Karl Marx Str. 12, Ufa, Bashkortostan, 450000 , Russia; elita@vmk.ugatu.ac.ru • Vadim M. Kartak; Ufa State Aviation Technical University, Karl Marx Str. 12, Ufa, Bashkortostan, 450000 , Russia; kvmail@mail.ru • Lidiya I. Vasilyeva; Ufa State Aviation Technical University, Karl Marx Str. 12, Ufa, Bashkortostan, 450000 , Russia; lidav@mail.ru # Health Care Quality & Long-Term Care Session: SD22 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Health Applications Section Track: Cluster: Room: Chair: Sandra Potthoff Chair Address: University of Minnesota, Carlson Sch. of Mgmt., Rm. 3-140, 321 19th Ave. South, Minneapolis, MN 55455-9940 Chair E-mail: potth001@tc.umn.edu Chair: Chair Address: Chair E-mail: SD22.1 Home Web Use to Support Traumatic Brain Injury Family Caregivers & Patients • Armando James Rotondi; University of Pittsburgh, 646C Scaife Hall, CCM Division, 200 Lothrop St., Pittsburgh, PA 15213; rotondi+@pitt.edu We will present data from a randomized evaluation of WE CARE (Web Enabled Caregiver Access to Information and Education) which is an interactive Web site designed to support those who are providing home care to a family member with a severe traumatic brain injury. SD22.2 withdrawn - author request of 10/10 • Sharon Schweikhart; Ohio State University, Hlth. Svcs. Mgmt. & Policy Div, Sch. of Pub. Hlth., 1583 Perry, Columbus, OH 43210-1234; schweikhart.1@osu.edu • Susan Meyer; University of Minnesota, Carlson Sch. of Mgmt., Operations & MS Dept., Minneapolis, MN 55455; smeyer@csom.umn.edu SD22.3 Preventing Human Immunodeficiency Virus Infections in Injection Drug Users: Allocationg Resources among Interventions & IDUs • Amy R. Wilson; University of California, Dept. of IE&OR, 4135 Etcheverry Hall, Berkeley, CA 94720; amy@ieor.berkeley.edu • James G. Kahn; University of California, Inst. Health Policy Studies, Box 0936, San Francisco, CA 94143; jgkahn@itsa.ucsf.edu • Shmuel S. Oren; University of California, Dept. of IE&OR, 4135 Etcheverry Hall, Berkeley, CA 94720; oren@ieor.berkeley.edu We discuss costs and benefits of targeting HIV prevention programs to IDU sub-populations, e.g., high-risk. Benefits depend on the definition of the sub-population. We develop a cost function that increases as the target population size decreases. We show how the optimal targeting and intervention policies vary with targeting costs. SD22.4 The Relationship between Quality Management Practices & Performance in Long-Term Care Facilities • Sandra Potthoff; University of Minnesota, Carlson Sch. of Mgmt., Rm. 3-140, 321 19th Ave. South, Minneapolis, MN 55455-9940; potth001@tc.umn.edu • Douglas Olson; University of Minnesota, Carlson Sch. of Mgmt., Rm. 3-140, 321 19th Ave South, Minneapolis, MN 55455-9940; douglas.m.olson-2@tc.umn.edu • John C. Anderson; University of Minnesota, Carlson Sch. of Mgmt., Rm. 3-140, 321 19th Ave South, Minneapolis, MN 55455-9940; janderson@csom.umn.edu • Robert L. Kane; School of Public Health, Box 197, 420 Delaware St. SE, Minneapolis, MN 55455; kanex001@maroon.tc.umn.edu • April Todd-Malmlov; University of Minnesota, Carlson Sch. of Mgmt., Rm. 3-140, 321 19th Ave South, Minneapolis, MN 55455-9940; todd0049@tc.umn.edu SEM is used to analyze survey data from 75 Minnesota nursing facilities to assess the relationship between quality management practices and resident satisfaction, employee satisfaction and financial performance. # Tutorial: Statistical Analysis of Simulation Output Session: SD23 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Room: Chair: David Goldsman Chair Address: Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332 Chair E-mail: sman@isye.gatech.edu Chair: Chair Address: Chair E-mail: SD23.1 Tutorial: Statistical Analysis of Simulation Output • David Goldsman; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332; sman@isye.gatech.edu The analysis of simulation output data is difficult - such output rarely satisfies the independence and normality assumptions one finds in standard statistics texts. We outline a number of ways around these problems. We also demonstrate these techniques on a number of real-world examples. # Application of Simulation in Material Handling Systems Design Session: SD24 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Material Handling Room: Chair: Douglas D. Gemmill Chair Address: Iowa State University, Dept. of IMSE, Ames, IA 50011 Chair E-mail: Chair: Chair Address: Chair E-mail: SD24.1 Simulation of a Shop Floor Logistics System • Michael H. Cole; University of Arkansas, 4207 Bell Eng. Ctr., Dept. of IE, Fayetteville, AR 72701; • Terry R. Collins; University of Arkansas, , AR; We will utilize simulation programming to further refine a multi-period model for the design and analysis of focused storage systems. The multi-period model will improve on past research in focused storage system designs by considering the dynamic nature of component demand. The new model will allow for multiple material handling modes, each with associated costs... SD24.2 Development of a Simulation Module for Controlled Spacing Induction of Packages from Multiple Conveyor Lines • Taimour El-Cheikh; University of Louisville, Dept. of IE, Louisville, KY 40292; • M. R. Wilhelm; University of Louisville, Dept. of IE, Louisville, KY 40292; We address the issue of designing induction conveyor systems by presenting the results of a study to design an induction sub-system simulator module. The characteristic events and variables of a typical induction conveyor control system are presented, as are the features of the simulation module developed. SD24.3 Closed-Loop Overhead Conveyors in Two-Stage Automated Production Flow Lines We discuss the application of computer simulation to the design of a closed-loop overhead monorail conveyor, which serves as buffer storage in a 2-stage automatic transfer line. We describe models to be used in predicting the various factors contributing to the inefficiency of the production flow line. SD24.4 Container Handling Physical Process Reengineering in Maritime Terminals • Ardavan Asef-Vaziri; Marquette University, Coll. of Bus. Admin., Milwaukee; • Behrokh Khoshnevis; University of Southern California, Dept. of IE/ES, Los Angeles, CA 90089-0193; We analyze the impact of instituting AS/RS, AGVS, a loop-based yard transportation network and a JIT inventory philosophy on the operations of maritime container terminals. # Sensitivity Analysis in Convex Optimization Session: SD25 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Optimization Section Track: Cluster: Nonlinear Programming Room: Chair: Katya Scheinberg Chair Address: IBM TJ Watson Research Center, PO Box 218, Rte. 134, Yorktown Heights, NY 10598 Chair E-mail: Chair: Chair Address: Chair E-mail: SD25.1 Perturbations of a Convex-Conic Program & Applications • Miguel S. Lobo; Stanford University; We consider how changes in the parameters of a convex-conic program translate into changes in the optimal value of the program objective. If the optimal value is convex with respect to the parameters of interest, we obtain a subgradient that is a bi-linear function of the optimal primal and dual variables... SD25.2 On the Sensitivity of Central Solutions in SDP • Shuzhong Zhang; Chinese University of Hong Kong, Dept. of SEEM, Shatin NT, , Hong Kong; zhang@se.cuhk.edu.hk • Jos F. Sturm; ; We will discuss properties of the analytic central path of an SDP problem under the perturbation of the problem data. Various results are proven under the strictly complementary condition. SD25.3 An Interior-Point Approach to Sensitivity Analysis We present our ideas on incorporating interior-point methods into sensitivity analysis in LP and SDP. We propose an interior-point approach and show that it bears a nice relationship with the approach based on maintaining the optimal partition in LP and SDP, which seems to be a natural way of performing sensitivity analysis. SD25.4 Classification of Self-Scaled Barrier Functions Self-scaled barrier functions are an important tool in the study of conic linear programming problems over symmetric cones. They are also closely related to Euclidean Jordan algebras. First we show that every self-scaled barrier function defines a Euclidean Jordan algebra. Then we use this algebra to obtain a complete classification of self-scaled barrier functions... # Stochastic Decision Processes Session: SD26 Date/Time: Sunday 15:00-16:30 Type: Invited Sponsor: Track: Cluster: Stochastic Models & Applications Room: Chair: Emmanuel Fernandez Chair Address: University of Arizona, SIE Dept., Tucson, AZ 85721-0020 Chair E-mail: emmanuel@sie.arizona.edu Chair: Chair Address: Chair E-mail: SD26.1 Risk-Sensitive Inventory Control Problems Motivated by the risk and variability introduced in the supply chain by short-term demand and market dynamics forecasts, we study an inventory control problem, under a stochastic demand process and with a risk (i.e., variance) sensitive optimality criterion. SD26.2 Modularity & Monotone Solutions in Risk-Sensitive Stochastic Decision Processes We consider controlled Markov chains models with a countable state space, under (exponential) total and discounted risk-sensitive cost criteria. A set of general conditions is presented to obtain monotonicity properties of value functions and optimal policies, based on submodularity arguments SD26.3 Partially Observable Markov Decision Processes with Average Cost • Aristotle Arapostahis; University of Texas, Electric. & Comp. Eng. Dept., Austin, TX 78712-1084; ari@mail.utexas.edu We present a new set of sufficient conditions for the existence of optimal stationary policies for the ergodic control problem in partially observed Markovian systems. Results are applicable to risk-free and risk-sensitive controlled Markov chains as well as 0-sum stochastic games. A comparative review of existing results in the literature and how this work improves on them is also presented... SD26.4 Inventory Allocation at a Semiconductor Company: Modeling & Optimization • Alexander O. Brown; Vanderbilt University, Owen Grad. Sch. of Mgmt., 401 21st Ave. South, Nashville, TN 37203; alex.brown@owen.vanderbilt.edu • Grace Lin; IBM Corporation, TJ Watson Research Ctr., PO Box 218, Yorktown Heights, NY 10598; gracelin@us.ibm.com • Markus Ettl; IBM, TJ Watson Research Ctr., PO Box 218, Rte. 134, Yorktown Heights, NY 10598; msettl@us.ibm.com • David D. Yao; Columbia University, IEOR Dept., New York, NY 10027-6699; yao@ieor.columbia.edu We develop a multi-echelon, multi-item inventory allocation model for the use in the semiconductor industry. The problem is formulated as a constrained nonlinear program, so as to find the safety stock targets that optimizes a system-wide service measure subject to constraints on allowable inventory holding cost by echelon. We develop an efficient algorithm that finds the optimal solution... # Negotiation to Win-Win Session: SD27 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Women in OR/MS Track: Cluster: Room: Chair: Robin Lougee-Heimer Chair Address: IBM TJ Watson Research Center, PO Box 218, Rte. 134, Yorktown Heights, NY 10598-0218 Chair E-mail: robinlh@us.ibm.com Chair: Chair Address: Chair E-mail: SD27.1 Negotiating Win-Win Everything is negotiable. This interactive talk will focus on creating win-win solutions to everyday workplace challenges using the PIA method of negotiating. The presentation material is based on feedback from a pre-conference, internet survey of INFORMS members. Join us for a fun and skill-enhancing presentation. # What Defines the Success of OR Applications in the Railroad Industry II Session: SD28 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: RASIG Track: Cluster: Room: Chair: Ingrid Schultze Chair Address: Reebie Associates Chair E-mail: ischultze@reebie.com Chair: Chair Address: Chair E-mail: SD28.1 What Defines the Success of OR Applications in the Railroad Industry II • Carl D. Martland; MIT, Dept. of Civil & Environ. Eng., Rm. 1-153, Cambridge, MA 02139; • Joe Bryan; Reebie Associates; • Martha Lawrence; T&MC; • Richard Gray; Union Pacific; Carl Martland: Preliminary RemarksC - Railroad Costing - Moderator to be announced Consultant Perspective - Joe BryanIndustry Perspective - To be announced International Railroad Perspective - Martha LawrenceD - Terminal Capacity & PlanningCarl Martland, Moderator: Overview of How We MeasureModeling Perspective - Carl Martland, MIT Industry Perspective - Richard Gray, Union Pacific # Modeling Generation & Transmission Planning in Electricity Markets Session: SD29 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: ENRE Track: Cluster: Room: Chair: Scott Rogers Chair Address: University of Toronto, Dept. of MIE, 5 Kings College Rd., Toronto, Ontario, M5S 3G8 , Canada Chair E-mail: rogers@mie.utoronto.ca Chair: Chair Address: Chair E-mail: SD29.1 Coalition Formation Methods in Transmission Expansion Planning: Bilateral Shapley Value & Kernel Approaches • J. Contreras; Universidad de Castilla La Mancha, 13071 Ciudad Real, La Mancha, , Spain; javier@ind-cr.uclm.es We present a multi-agent-based system that provides a cooperation plan and its associated cost allocation scheme to assist transmission expansion planners in their decision making process. Among several coalition formation and cost allocation criteria, we have selected the bilateral Shapley value and the kernel approaches as theoretical foundations. SD29.2 Modeling Inflow Uncertainty in Electricity Markets using Dynamic Programming & Complementarity Problem Techniques • Mariano Ventosa; Universidad Pontificia Comillas, Inst. de Investigacion Tec., Alberto Aguilera 23, Madrid, 28015 , Spain; mariano.ventosa@iit.upco.es • Antonio Garcia-Alcalde; Universidad Pontificia Comillas, Inst. de Investigacion Tec., Alberto Aguilera 23, Madrid, 28015 , Spain; • Michel Rivier; Universidad Pontificia Comillas, Inst. de Investigacion Tec., Alberto Aguilera 23, Madrid, 28015 , Spain; • Andres Ramos; Universidad Pontificia Comillas, Inst. de Investigacion Tec., Alberto Aguilera 23, Madrid, 28015 , Spain; We propose a new approach for addressing the long-term hydro-thermal coordination of an electricity generation company operating in a competitive market and inflow uncertainty context. The methodology employs the traditional SDP techniques, being the sub-models at each SDP stage stated in terms of a mixed complementarity problem in order to represent the electricity market equilibrium. SD29.3 Private & Public Transmission Ownership & Investment • Amy B. Craft; US Department of Justice, Antitrust Division, 600 East St. NW, Ste. 10-011, Washington, DC 20530; amy.craft@usdoj.gov We discuss the long-run efficiency of various transmission ownership scenarios. Market power among energy service providers distorts nodal prices as a measure of transmission capacity's social value. Conversely, establishing optimal transmission ownership policy may curb this strategic behavior. This has particular implications for the design of regional transmission organizations. SD29.4 GENCOMP 3.0: A Model of Competition among Generating Firms • Scott Rogers; University of Toronto, Dept. of MIE, 5 Kings College Rd., Toronto, Ontario, M5S 3G8 , Canada; rogers@mie.utoronto.ca We describe a model of competition among electric generating firms. It allows different types of firms: strategic, traditional and supply curve and various classes of demand. It solves for a Nash equilibrium. Applications to real systems are described. # Web-Based Tactical Applications of OR II Session: SD30 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Computing Society Track: Cluster: Room: Chair: Radhika Kulkarni Chair Address: SAS Institute Inc., SAS Campus Dr., Cary, NC 27513 Chair E-mail: radhika.kulkarni@sas.com Chair: Chair Address: Chair E-mail: SD30.1 Improving the Effectiveness of Testing in the Manufacture of Computers in a Make-to-Order System • Suheil Nassar; Personal Systems Group, IBM, RTP, NC; nassar@us.ibm.com • Mike Newton; Personal Systems Group, IBM, RTP, NC; • Linda Hunke; University of North Carolina, Dept. of OR, Chapel Hill, NC 27599-3180; • Vidyadhar G. Kulkarni; University of North Carolina, Dept. of OR, Smith Bldg., CB 3180, Chapel Hill, NC 27599-3180; vkulkarni@email.unc.edu • Michelle Opp; University of North Carolina, Dept. of OR, Chapel Hill, NC 27599-3180; • J. Scott Provan; University of North Carolina, Dept. of OR, Chapel Hill, NC 27599-3180; • Shaler Stidham, Jr.; University of North Carolina, Dept. of OR, Chapel Hill, NC 27599; We consider a computer manufacturing operation that includes a significant performance-testing component. We develop an integer programming formulation to decide the optimal set of tests to conduct so that the out-of-box quality level is maximized subject to a lower bound on the throughput. A detailed simulation model is developed to validate the optimal testing policies produced by the integer program. SD30.2 Web-Basesd Collaborative Forecasting • Farah Marasigan; Supply Chain Consultants, Inc., 5460 Fairmont Dr., Linden Green Ctr., Wilmington, DE 19808; fmarasigan@supplychain.com • Greg Werker; Supply Chain Consultants, Inc., 5460 Fairmont Dr., Linden Green Ctr., Wilmington, DE 19808; gwerker@supplychain.com Collaborative forecasting involves collecting and reconciling information from diverse sources to produce a unified statement of demand. The major technical challenge is reconciling different views of market demand. Most businesses have 2 organizations dealing with marketplace demand - sales views demand by customer and marketing views by product. Each group has several individuals managing the demand of their customers or products... SD30.3 Cycle Time Reduction on the Web at Lockheed Martin Lockheed Martin's Astronautics Division has been using OR tools for several web-based applications to monitor the performance of the production systems used for building the Titan and Atlas missiles. We present a brief overview of these applications and describe some of the methods used in the Cycle Time Reduction project. # OR Modeling in Engineering Designs: Lectures in Honor of Carl M. Harris Session: SD31 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Computing Society Track: Cluster: Room: Chair: Karla Hoffman Chair Address: George Mason University, Systems Eng. & OR Dept., 4400 University Dr., MS 4A6, Fairfax, VA 22030 Chair E-mail: khoffman@gmu.edu,, http://iris.gmu.edu/~khoffman Chair: Chair Address: Chair E-mail: SD31.1 Optimal Reliability Allocation • Nozer D. Singpurwalla; George Washington University, Sch. of Eng. & Applied Sci., 707 22nd St. NW, Washington, DC 20052; • Jim Falk; George Washington University, Sch. of Eng. & Applied Sci., 707 22nd St. NW, Washington, DC 20052; • Yefim Vladimirsk; George Washington University, Sch. of Eng. & Applied Sci., 707 22nd St. NW, Washington, DC 20052; An important issue in engineering design is reliability allocation leading to an interesting interplay between distribution theory and optimization. Each optimization problem boils down to finding the fixed points of a function within a unit hypercube. Crossing properties of star-ordered distribution functions are used to establish the existence of the fixed points. SD31.2 Applying Predictive Dialing Methods to Real-Time Adaptive Control of Telecommunications Systems • Douglas A. Samuelson; InfoLogix, Inc., 8711 Chippendale Ct., Annandale, VA 22003; Predictive dialing, used in outbound telephone call centers, is a special case of controlling queuing systems that initiate attempts to acquire customers. The real-time version of the solution offers considerable promise for controlling telecommunications systems for which conventional steady-state queueing models are intractable. SD31.3 Time Critical Targeting using Bayesian Network Analysis • Mark Frymire; Science Applications International Corp., Analysis & Support Div., 1710 SAIC Dr., MS 3-6-1, McLean, VA 22102; We present a complicated probability estimation problem coupled with Bayesian decision analysis in order to determine whether to attack a given mobile, high-value enemy ground target, based on the timing of intelligence obtained and the weapon's capabilities. The model is used to improve the efficiency of the difficult task of attacking transient targets. # Political Institutions & Firm Strategy Session: SD32 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Organization Science Section Track: Cluster: Room: Chair: Brian Silverman Chair Address: Harvard Business School, 243 Morgan Hall, Soldiers Field, Boston, MA 02163 Chair E-mail: bsilverman@hbs.edu Chair: Chair Address: Chair E-mail: SD32.1 Organizational, Imitative & Network Learning of Political Hazards: Market Entry Strategies of Japanese Multinational Corps., 1990-97 • Witold Henisz; University of Pennsylvania, The Wharton Sch., 2021 Steinberg Hall, Philadelphia, PA 19104-6370; henisz@wharton.upenn.edu • Andrew Delios; , Hong Kong, , Hong Kong; We perform a longitudinal examination of the relationship between political hazards in a prospective host country and the market-entry mode chosen by multinational corporations. We demonstrate that political hazards are an important determinant of entry mode but also that this relationship is moderated by the information possessed by the firm. SD32.2 Modular Governance: Explaining the Effect of Information Technology on the Boundaries of the Firm • Jack A. Nickerson; Washington University, Olin Sch. of Bus., 1 Brookings Dr., CB 1133, St. Louis, MO 63130-4538; • Sergio G. Lazzarini; Washington University, Olin Sch. of Bus., 1 Brookings Dr., CB 1133, St. Louis, MO 63130-4538; lazzarinis@mail.olin.wusl.edu Since new ITs reduce both external and internal transaction costs, their effect on firm boundaries is ambiguous. We propose a mechanism that explains unambiguously trends in vertical disintegration: increases in governance modularity due to new modular ITs. Modularity facilitates participation of multiple firms in a transaction through standardized interfaces, reducing expropriation problems even when asset-specificity is present. SD32.3 A Positive Political Theory of Internet Corporate for Assigned Names & Numbers We present a positive policy theory of the creation of ICANN and its policy for resolving domain names disputes. We suggest that while certain interests failed in advancing their interests regarding creation of domain name policy, they succeeded in designing ICANN to secure their policy interests in the execution of the disputed policy. SD32.4 Business Groups & Risk Sharing around the World • Tarun Khanna; Harvard Business School, Morgan Hall 243, Soldiers Field, Boston, MA 02163; • Yishay Yafeh; Hebrew University, Jerusalem, , Israel; We investigate risk sharing by diversified business groups in 16 economies. We find that groups smooth members' income, but this effects magnitude is small. We find no correlation between extent of income smoothing and capital market development. This suggests that risk sharing is not the primary reason for ubiquity of business groups. We do, however, find evidence of substantial 'liquidity smoothing' in one country. # Production Planning III Session: SD33 Date/Time: Sunday 15:00-16:30 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Timothy J. Flynn Chair Address: University of Alabama, Commerce & Bus. Admin. Coll., Box 870223, Hoover, AL 35487-0023 Chair E-mail: timjflynn99@aol.com,, flynnt@proctr.cba.ua.edu Chair: Chair Address: Chair E-mail: SD33.1 withdrawn - author request of 9/18 SD33.2 Developing a Line Balancing Model of Product Costing for Manufacturing Industries • Massimo de Falco; University of Lecce, Via Arnesano, Lecce, 73100 , Italy; defalco@unina.it • Maria Elena Nenni; University of Naples, Piazzale Tecchio 80, Naples, 80125 , Italy; nenni@uniroma2.it We illustrate an evolutionary model of 'product costing' focused on short- and medium-term control of idle capacity and line balancing. The model is subsequently analyzed in terms of its impact on production planning and on overall management. The model is validated with an application in a chemical industry. SD33.3 withdrawn - author request of 10/28 SD33.4 Simulated Annealing for Mixed-Model Assembly Line Sequencing Mixed-model assembly implies workstations with variable processing times. In some lean environments, the processing times are also sequence dependent. A characteristic of dominant sequences for such a case is used to construct an efficient neighborhood structure. A simulated annealing algorithm and experimental results are presented. # Decision Analysis IV Session: SD34 Date/Time: Sunday 15:00-16:30 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Robert W. Lawler Chair Address: Boeing, 17513 155th Ave. SE, Renton, WA 98058 Chair E-mail: robert.w.lawler@boeing.com Chair: Chair Address: Chair E-mail: SD34.1 Supply Contracts & Quality Choice in Food Chains • Severine Gaucher; INRA ESR/CIRAD TERA, 65 Bd de Brandebourg, Ivry Sur Seine, 94205 , France; gaucher@ivry.inra.fr • Louis Georges Soler; INRA ESR, 65 Bd de Brandebourg, Ivry Sur Seine, 94205 , France; soler@ivry.inra.fr We discuss ordering policies and their impact at the level of product quality in agro food chains. We show that without an early buyer's commitment, high quality production may be impossible. We then compare several flexibe contracts and their impact on the profits of the buyer and the supplier. SD34.2 Subjectivity in Human Decision-Making • T. S. Chidambaram; The Analytic Team, 6196 Oxon Hill Rd., Ste. 220, Oxon Hill, MD 20745; rajuateam@aol.com Human decision-making is not merely an intellectual activity but also involves subjective layers of personality. Further, it has an element of unpredictability that is conceptually similar to the notion of superposition in quantum mechanics. Against this background, we discuss the role of objective DSSs such as OR/MS in decision-making. SD34.3 A Decision-Analytic Approach for Prioritizing Second-Generation Molecules • Phillip Beccue; Amgen, 1 Amgen Center Dr., MS 27-5-B, Thousand Oaks, CA 91320; phil@beccue.com Prioritizing R&D projects often requires difficult tradeoffs among competing objectives. I will discuss a real application illustrating a rigorous, quantitative approach to selecting a preferred second-generation drug candidate. The project team used multi-attribute utility analysis to clarify the appropriate balance between safety, efficacy, convenience and cost. SD34.4 An E-Business Frame for Decision Analysis The transition from traditional business models to the business-to-business paradigm provides opportunities for improved financial returns, but introduces new legal and technical risks. Decision-makers often ignore these risks or delay decisions. We discuss issues that need incorporation into the frame and decision model. # Information Systems II Session: SD35 Date/Time: Sunday 15:00-16:30 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Pedro R. Kanof Chair Address: George Washington University, Sch. of Eng. & Applied Sci., 1669 32nd St. NW, Washington, DC 20007 Chair E-mail: pkanof@aol.com Chair: Chair Address: Chair E-mail: SD35.1 Information Technology Infrastructure as a Strategic upon in Top Container Ports • Rafael M. Oliveira; Fundacao Universidade Federal do Rio Grande, Gomes Freire 769, Rio Grande, RS 96200-470 , Brazil; rmelloo@rgd.terra.com.br • Antonio G. Macada; Fundacao Universidade Federal do Rio Grande, Silva Paes 369/702, Rio Grande, RS 96200-340 , Brazil; acgmacada@adm.ufrgs.br The IS literature shows the information technology infrastructure as a strategic upon. We investigate issues related to how IT is being used in top container ports and if the infrastructure differences among them result in competitive advantage. SD35.2 Evaluating the Impact of Large-Scale Information Systems • Zeynep Onay; Middle East Technical University, Dept. of Bus. Admin., Inonu Bulvari, Ankara, 06531 , Turkey; onay@ba.metu.edu.tr • Cemal Akyel; Middle East Technical University, Dept. of Bus. Admin., Inonu Bulvari, Ankara, 06531 , Turkey; akyel@ba.metu.edu.tr We present a 3-dimensional framework for assessing the effectiveness of the National Health Information System developed by the Ministry of Health of the Republic of Turkey. The framework constitutes a foundation for developing a comprehensive model for evaluating large-scale public information systems. SD35.3 Reusable Software Patterns for E-Business Applications A software pattern is a parameterized collaboration of classes. It can capture a repeatably usable design of a component of a business model. We present some patterns developed for e-business applications and discuss research issues in this field. SD35.4 Planning E-Business Projects • Pedro R. Kanof; George Washington University, Sch. of Eng. & Applied Sci., 1669 32nd St. NW, Washington, DC 20007; pkanof@aol.com I focus on the modules that are required in an E-business project and defined their functions and interelationships. I discuss ways to access the market by exploring alternatives for each module and conclude by describing the process for planning and controlling the implementation phases of the solution. # Logistics Management II Session: SD36 Date/Time: Sunday 15:00-16:30 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Chulung Lee Chair Address: University of Waterloo, 200 University Ave. West, CPH4347-200, Waterloo, Ontario, N2L 3G1 , Canada Chair E-mail: chulung@engmail.uwaterloo.ca Chair: Chair Address: Chair E-mail: SD36.1 Solving the Production-Distribution System Design Problem using an Interior-Point-Cutting-Plane Method We propose a 2-level Lagrangean relaxation approach for the production-distribution system design problem, where master problems are solved using the analytic-center cutting-plane method (ACCPM). At the first level, ACCPM is used to solve the dual Lagrangean problem, while feasible solutions are generated using a primal heuristic. At the second level, ACCPM is incorporated into a B&P scheme. SD36.2 Hierarchical Logistics & Manufacturing Frameworks • Tan Miller; Warner-Lambert Co., 201 Tabor Rd., Morris Plains, NJ 07950; tan.miller@wl.com We illustrate how standard OR methodologies such as optimization and simulation can be combined with qualitative information and spreadsheet-typed quantitative analyses to evaluate manufacturing and distribution network problems. We reviewed a sample of hierarchical manufacturing, distribution and transportation network design approaches used in real world applications. SD36.3 A Simple Heuristic for Dynamic Order Sizing & Supplier Selection with Time-Varying Data We consider the problem of supplier selection and purchase ordersizing for a single item under dynamic demand conditions. Suppliersoffer quantity discounts which may vary over time. A new model formulation for this problem is developed and a simple but easilyextendible heuristic procedure is presented and tested. SD36.4 An Empty Trailer Repositioning Model Intermodal freight movements often result in the accumulation of trailers in some markets while others experience shortages. An LP model with user-friendly interface is used at Schneider National to determine daily repositioning of trailers from surplus markets to deficit markets with the objective of maximizing net 'value' while satisfying customer service requirements. SD36.5 A Study on Product Locations for a One-Dimensional Storage Rack Order picking is probably the most important at distribution centers. Increasing throughput, pick accuracy and customer service is important while reducing the cost. Among factors that affect the system performance, storage locations are the most critical. We solve several problems of determining storage locations in a one-dimensional storage rack. # Optimization Techniques IV Session: SD37 Date/Time: Sunday 15:00-16:30 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Urmila Diwekar Chair Address: Carnegie Mellon University, Baker Hall 129, Dept. of EPP, 5000 Forbes Ave., Pittsburgh, PA 15213 Chair E-mail: ud01@andrew.cmu.edu Chair: Chair Address: Chair E-mail: SD37.1 A Support Vector Machine for Regression & Applications to Financial Forecasting • Huseyin Ince; University of Oklahoma, Sch. of IE, 202 West Boyd., Rm. 124, Norman, OK 73019-0631; ince@ou.edu • Theodore B. Trafalis; University of Oklahoma, Sch. of IE, 202 West Boyd, Rm. 124, Norman, OK 73019; trafalis@ou.edu We compare the SVM developed by Vapnik with other techniques such as back propagation and radial basis function networks for financial forecasting applications. The theory of the SVM algorithm is based on statistical learning theory. Training of SVMs leads to a quadratic programming problem. Preliminary computational results for stock price prediction are also presented. SD37.2 Experimental Optimization in Flat Hot & Cold Rolling • Mohamed H. Gadallah; Cairo University, Statistical Studies & Research, Giza, Cairo, 12613 , Egypt; m_gadallah@hotmail.com We offer a design methodology to integrate robust and experimental techniques with the flat hot and cold rolling process. Three major control parameters are considered: geometric, frictional and material parameters. The methodology is applied to process spread and out-of-straightness tolerances. An equivalent table showing maximum deviation range vs. process conditions is also produced. SD37.3 A Network-Based Approach to the General Equal Flow Problem • Herminia I. Calvete; Universidad de Zaragoza, Dept. de Metodos Estadisticos, Pedro Cerbuna 12, Zaragoza, 50009 , Spain; herminia@posta.unizar.es A simplex algorithm is developed for the general equal flow problem, a MCNF problem with additional side constraints requiring that the flow of arcs in some given sets of arcs takes on the same value. It exploits the network structure and requires only slight modifications of the network simplex algorithm. SD37.4 withdrawn - author request of 10/19 • Jaya P. Moily; University of Baltimore, 1420 North Charles St., Merrick Sch. of Bus., Baltimore, MD 21201-5779; jmoily@ubmail.ubalt.edu # Industry Applications II Session: SD38 Date/Time: Sunday 15:00-16:30 Type: Contributed Sponsor: Track: Cluster: Room: Chair: David S. Kim Chair Address: Oregon State University, Dept. of IME, Corvell Hall, Corvallis, OR 97331-2407 Chair E-mail: odomreggie@aol.com Chair: Chair Address: Chair E-mail: SD38.1 Analysis of Maximum Flow Algorithms for the Optimal Contours of a Mine • Michel Gamache; Ecole Polytechnique de Montreal, CP 6079 succ. Center-Ville, Montreal, Quebec, H3C 3A7 , Canada; michel.gamache@courriel.polymtl.ca • Genevieve Auger; Ecole Polytechnique de Montreal, CP 6079 succ. Center-Ville, Montreal, Quebec, H3C 3A7 , Canada; The operating schedule problem in an open pit mine can be associated with the problem of maximal closure on a graph with resource constraints. The Lagrangian relaxation represents the most interesting approach to solve this problem. The subproblem resulting from the relaxation of the resources constraints is the classic maximal closure problem, which is solved by a maximal flow algorithm... SD38.2 Overcoming Technical Challenges in Customized Enterprise Resource Planning/Advances Planning & Scheduling Data Integration We will highlight the technical challenges faced in developing customized interfaces between ERP (e.g. SAP RL/3) and APS systems (e.g. i2 SCP). One of the most critical elements of integrating the 2 systems is an early development of routines and processes to measure and improve data quality in the ERP database. We will discuss how prevalent designs rate in their robustness... SD38.3 Effective Heuristics for Multi-Product Partial Shipment Models We consider 3 shipment models, motivated by real applications where multiple customers that ask for sets of products are satisfied from inventory in the best manner possible. The restrictions posed by the customers are that the first shipment be at least a minimum fraction of the total demand, that the second shipment (if any) should not be too small nor be further split... SD38.4 Operations Research in an Engineering & Development Organization • David S. Kim; Oregon State University, Dept. of IME, Corvell Hall, Corvallis, OR 97331-2407; odomreggie@aol.com Applications of OR to decisions in the operation of an engineering & development organization is relatively new. We give an overview of various IE/OR problems/opportunities encountered at a large engineering organization. # Quality Management & Statistics I Session: SD39 Date/Time: Sunday 15:00-16:30 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Danny I. Cho Chair Address: Brock University, 500 Glenridge Ave., St. Catharines, ON, L2S 3A1 , Canada Chair E-mail: dcho@taro.bus.brocku.ca Chair: Chair Address: Chair E-mail: SD39.1 Economic Design of QC Charts for a Production Line We consider the economic design of QC charts for a group of machines (on a production line) facing random opportunities of inspection line stoppages). A single machine case is developed exactly. An approximate scheme is suggested for grouping machines as opportunity-takers and nontakers. Comparison with classic methods is presented. SD39.2 Maximum Entropy Dirichlet Modeling of the Choice of a Long Distance Provider • Ehsan S. Soofi; University of Wisconsin, Sch. of Bus. Admin., PO Box 742, Milwaukee, WI 53201; esoofi@uwm.edu,, http://www.umw.edu/~esoofi/ • Thomas A. Mazzuchi; George Washington University, Washington, DC 20052; • Refik Soyer; George Washington University, Monroe Hall 403, MS Dept., 2115 G St. NW, Washington, DC 20052; soyer@gwu.edu • Joseph J. Retzer; Maritz Marketing Research, 1415 West 22nd St., Ste. 800, Oak Brook, IL 60523; We use MED procedure to model consumer choice of the long distance provider. The MED is a computer-intensive method that uses Dirichlet prior and various attribute constraints and produces loglinear and logit models, prior and posterior distributions for the model parameters and for the Kullback-Leibler information of the model fit. SD39.3 Information Technology, Reengineering & TQM: Three Competitive Strategies for Success in Today's Business • Danny I. Cho; Brock University, 500 Glenridge Ave., St. Catharines, ON, L2S 3A1 , Canada; dcho@taro.bus.brocku.ca IT, reengineering and TQM have changed our ways of doing business in this competitive, global business world. We study the strategic implications of using each of these fundamental but important business strategies. Secondly, we examine any synergies that an organization might foresee when applying any combination of these strategies. Finally, we identify several challenges that business managers might face in implementing these strategies. # Stochastic Processes Session: SD40 Date/Time: Sunday 15:00-16:30 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Debra A. Elkins Chair Address: General Motors R&D Center, 30500 Mound Rd., MC 480-106-359, Box 9055, Warren, MI 48090-9055 Chair E-mail: debra.elkins@gm.com Chair: Chair Address: Chair E-mail: SD40.1 no show SD40.2 Bayesian Analyses of Nonhomogeneous Markov Chains We present a hierarchical Bayesian approach for analyses of time non-homogeneous Markov chains. The Markov chain model is used for describing transition behavior of emotionally disturbed children in a treatment program. We consider extensions of the model where the transition probabilities are functions of covariates as well as time. The developed methodology is applied to some real data. SD40.3 Scheduling & Sequencing Customers in an Appointment System An effective algorithm for optimizing the schedule of customers in an appointment system is presented. Service distributions are general and distinct. Cost is defined as a convex combination of customer waiting times and server idle time. Heuristics for sequencing customers are also discussed. Applications include JIT and medical appointment systems. SD40.4 A Parallel Algorithm for Computing the Markov Renewal Kernel • Debra A. Elkins; General Motors R&D Center, 30500 Mound Rd., MC 480-106-359, Box 9055, Warren, MI 48090-9055; debra.elkins@gm.com • M. A. Wortman; Texas A&M University, 238 Zachry Eng. Ctr., College Station, TX 77843-3131; wortman@acs.tamu.edu We introduce a parallel algorithm for computing tight upper and lower bounds on the Markov renewal kernel. Our numerical results allow study of transient behavior of Markov renewal and semi-regenerative processes. Complexity analysis shows the algorithm to be of order O(k/2) speedup, when k processors are used. # The Daniel H. Wagner Prize for Excellence in OR Session: SD41 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: CPMS Track: Cluster: Room: Chair: Joseph H. Discenza Chair Address: SmartCrane, Inc., a Daniel H. Wagner Associates Company, 2 Eaton St., Ste. 500, Hampton, VA 23669 Chair E-mail: joeh@discenza.com Chair: Chair Address: Chair E-mail: SD41.1 Saving Tax Dollars for the Federal Government: Offsite Federal Employee Site Selection Software SD41.2 Rebuilding the Coal Model in the Energy Information Administration's National Energy Modeling System • Melinda Hobbs; Energy Information Administration, 1000 Independence Ave. SW, Washington, DC 20585; • Michael Mellish; Energy Information Administration, 1000 Independence Ave. SW, Washington, DC 20585; • Frederic H. Murphy; Temple University, Sch. of Bus. & Mgmt., Philadelphia, PA 19122; v5256e@vm.temple.edu • Richard Newcombe; Energy Information Administration, 1000 Independence Ave. SW, Washington, DC 20585; • Reginald C. Sanders; OnLocation, 8100 Oak St., Ste. 300, Arlington, VA 22027; rsanders@onlocationinc.com • Peter C. Whitman; Pace International; SD41.3 Optimal Design of a Data Offload Network • John Shortle; Qwest Advanced Technologies, 4001 Discovery Dr., Boulder, CO 80303; jshortle@gmu.edu • Dennis C. Dietz; Qwest Advanced Technologies, 4001 Discover Dr., Boulder, CO 80303; dxdiet2@uswest.com • Paul A. Katz; Qwest Advanced Technologies, 4001 Discovery Dr., Boulder, CO 80303; pkatz@uswest.com • Craig Williamson; Qwest Advanced Technologies, 4001 Discover Dr., Boulder, CO 80303; cbwill3@uswest.com • Jim Koehler; Qwest Advanced Technologies, 4001 Discover Dr., Boulder, CO 80303; jkoehle@uswest.com • Amie Elcan; Qwest Advanced Technologies, 4001 Discovery Dr., Boulder, CO 80303; aelcan@uswest.com # Journal of Quality Technology Papers Presentations Session: SD42 Date/Time: Sunday 15:00-16:30 Type: Sponsor/Invite Sponsor: Quality, Statistics & Reliability Section Track: Cluster: Reliability & Quality Control Room: Chair: Robert L. Mason Chair Address: Southwest Research Institute, 6220 Culebra Rd., San Antonio, TX 78228-0510 Chair E-mail: rmason@swri.edu Chair: Chair Address: Chair E-mail: SD42.1 Non-Parametric Control Charts: An Overview & Some Results • Subha Chakraborti; University of Alabama; schakrab@cba.ua.edu • Van der Laan; ; • S. T. Bakir; University of Alabama; No abstract supplied. SD42.2 Some Properties of EWMA Feedback Quality Adjustment Schemes for Drifting Disturbances • Enrique Del Castillo; Pennsylvania State University, Dept. of IME, 310 Leonhard Bldg., University Park, PA 16802; exd13@psu.edu No abstract supplied. # Research & Educational Issues in Management of Technology Session: SD43 Date/Time: Sunday 15:00-16:30 Type: Sponsored Sponsor: Technology Management Section Track: Cluster: Room: Chair: Michael K. Badawy Chair Address: Virginia Polytechnic Institute & State University, PO Box 2931, Merrifield, VA 22116-2931 Chair E-mail: mbadawy@vt.edu Chair: Chair Address: Chair E-mail: SD43.1 Integrating Distance Learning Technology in an M.S. MOT Program • William T. Flannery; University of Texas, Div. of Mgmt. & Mktg., 6900 North Loop, 1604 West, San Antonio, TX 78249; flannery@lonestar.utsa.edu • W. Austin Spivey; University of Texas, Div. of Mgmt. & Mktg., 6900 North Loop, 1604 West, San Antonio, TX 78249; wspivey@utsa.edu The M.S. in Management of Technology program at UTSA offers 'on-line' courses to meet the varying needs of its graduate students. We discuss our experiences in applying and integrating distance learning technologies to the task of developing and delivering the on-line program for remote students. SD43.2 Some Educational Issues in Engineering & Technology Management Programs • Dundar F. Kocaoglu; Portland State University, Engineering Mgmt. Program, Portland, OR 97207-0751; kocaoglu@emp.pdx.edu We explore some of the issues involved in the design, development and delivery of educational programs in the field of engineering and technology management. SD43.3 Managing Virtual Teams: An Educational Perspective In a distance learning environment, managing virtual teams must address the issues of: communication, team leadership, inclusion or isolation of members, team building, process flow and deliverables and technical and emotional support. Results of student surveys are presented evaluating these and other challenges to virtual team membership. SD43.4 Some Future Research Directions in Technology & Innovation Management • Michael K. Badawy; Virginia Polytechnic Institute & State University, PO Box 2931, Merrifield, VA 22116-2931; mbadawy@vt.edu We will assess the current status of technology and innovation management research and identify some paths for advancing the scholarly research in this field. # Software Demonstration II Session: SD45 Date/Time: Sunday 15:00-16:30 Type: Software Demo Sponsor: Track: Cluster: Room: Chair: Alkis Vazacopoulos Chair Address: Dash Optimization, Inc., 135 West 27th St., 7th Fl., New York, NY 10001 Chair E-mail: alkis@dashopt.com Chair: Barry MacKichan Chair Address: MacKichan Software, Inc. Chair E-mail: barry@mackichan.com SD45.1 XPRESS-MP Version 12 • Alkis Vazacopoulos; Dash Optimization, Inc., 135 West 27th St., 7th Fl., New York, NY 10001; alkis@dashopt.com We report on recent progress on XPRESS-MP, Dash's state-of-the-art software solution for modeling and optimization. Special attention will be given to the highlights of the new release 12, including model cuts, cut directives, quadratic programming, parallel barrier and parallel IP optimizers. SD45.2 Increasing Your Productivity with Scientific Workplace 3.5 This presentation will cover the basic uses of Scientific WorkPlace 3.5 to write for publication, to compute and plot using the included computer algebra system (either MuPAD or Maple) and to publish mathematics on the Web. We will show a large number of examples. # The Balanced Scorecard Session: MA01 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: Decision Analysis Society Track: Cluster: Room: Chair: James L. Ritchie-Dunham Chair Address: University of Texas, Dept. of MSIS, 3615 Aspen Creek Parkway, Austin, TX 78749 Chair E-mail: jimrd@sdsg.com Chair: Chair Address: Chair E-mail: MA01.1 The Theory behind the Balanced Scorecard • James L. Ritchie-Dunham; University of Texas, Dept. of MSIS, 3615 Aspen Creek Parkway, Austin, TX 78749; jimrd@sdsg.com The BSC presents an attractive concept for measuring and managing organizational performance. We present current theoretically ungrounded BSC 'best practice,' as well as relevant theories that could strengthen it, such as multiple criteria decision making, means-ends analysis, statistical cause-effect methods, system dynamics and ERP systems. MA01.2 Designing a Balanced Scorecard • Ralph L. Keeney; University of Southern California, Ctr. for Telecomm. Mgmt., 101 Lombard St., Ste. 704W, San Francisco, CA 94111; keeneyr@aol.com A BSC requires both a scorecard and balancing. Creating a scorecard requires selecting elements to score and measures to describe performance on elements. Balancing requires evaluating different performance levels on an element and equating levels across different elements. Value-focused thinking and multi-attribute utility provide conceptual foundations for these requirements. MA01.3 Influence Diagram Models for the Balanced Scorecard • Ross D. Shachter; Stanford University, Dept. of MS & Eng., Serra House, Stern Hall, Stanford, CA 94305-4026; shachter@stanford.edu The BSC approach to measuring and managing corporate strategy lacks a specific methodology for implementation. Influence diagrams bring a solid theoretic foundation to the representation of the structure and numbers that capture the causal relationships and dependencies needed to implement the BSC. MA01.4 The Dynamic Scorecard: Objectives, Policies & Resources • Hal Rabbino; SDSG LLC, 11915 Stone Hollow Rd. #1527, Austin, TX 78757; halr@sdsg.com • James L. Ritchie-Dunham; University of Texas, Dept. of MSIS, 3615 Aspen Creek Parkway, Austin, TX 78749; jimrd@sdsg.com The BSC 'operationalizes strategy,' suggesting a balanced policy set for achieving organizational goals over time. Developing policies about the accumulation and utilization of strategic resources over time requires dynamic thinking. This presents a forum for integrating periodic, value-focused decision making with policies about making decisions - DA meets SD. # Tutorial: Applying Advanced Integrated Real-Time DSSs to Help Airlines Move to the Next Level of Efficiency Session: MA02 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Room: Chair: Gang Yu Chair Address: University of Texas, Dept. of MSIS, Gras. Sch. of Bus. Admin., Austin, TX 78712-1175 Chair E-mail: yu@uts.cc.utexas.edu Chair: Chair Address: Chair E-mail: MA02.1 Tutorial: Applying Advanced Integrated Real-Time Decision Support Systems to Help Airlines Move to the Next Level of Efficiency • Gang Yu; University of Texas, Dept. of MSIS, Gras. Sch. of Bus. Admin., Austin, TX 78712-1175; yu@uts.cc.utexas.edu The airline's product is measured by price, flexible schedules, on-time performance, safety, satisfactory in-flight services, proper baggage handling and convenient ticket purchases. To meet these demands and to provide a high-quality and low-cost product, airlines spend a tremendous amount of resources, time and effort to generate profitable and cost-effective fare classes, flight schedules, fleet plans, aircraft routes, crew pairings, gate assignments, etc... # Models of Supply Chain Coordination Session: MA03 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: MSOM Track: Cluster: Room: Chair: Yunzeng Wang Chair Address: Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106-7235 Chair E-mail: yxw36@po.cwru.edu Chair: Chair Address: Chair E-mail: MA03.1 Supply Chain Contracting to Encourage Downstream Investment in Cost Reduction • Stephen M. Gilbert; University of Texas, CBA 4.202, B6300, Austin, TX 78712; steve.gilbert@bus.utexas.edu • Viswanath Cvsa; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106; vxc10@po.cwru.edu Firms in supply chains tend to benefit from investments in cost reduction that are made by their downstream partners. By making advance commitments to wholesale prices, an upstream firm can encourage such downstream investments by eliminating a 'hold-up' problem. We adopt the perspective of the upstream firm... MA03.2 Models of Channel Coordination with Asymmetric Demand Information • Apostolos N. Burnetas; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106-7235; atb4@po.cwru.edu • Stephen M. Gilbert; University of Texas, CBA 4.202, B6300, Austin, TX 78712; steve.gilbert@bus.utexas.edu • Craig E. Smith; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106; ces14@po.cwru.edu A wholesaler sells a product to a retailer. The distribution of the end customer demand is known to the retailer but partially known to the wholesaler. We develop quantity discount pricing policies that guarantee retailer incentive compatibility and maximize expected wholesaler or channel profits and discuss the benefits of coordination. MA03.3 Adverse Effects of Overreaction to Demand Changes & Improper Demand Forecasting in a Supply Chain Inventory managers tend to overreact to demand changes. Using a serially correlated demand model, we show that this overreaction leads to sub-optimal replenishment policies internally and causes an undue bullwhip effect externally. We also show that using an improper forecasting approach may cause similar adverse effects. MA03.4 Dynamic Competition with Inventory-Dependent-Demands • Yunzeng Wang; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106-7235; yxw36@po.cwru.edu We model the multi-period competition between 2 retailers where demand is inventory-dependent and customers may switch between retailers in case of a stock-out. We derive the equilibrium optimal inventory-policies and show that such a policy is myopic in nature and can be found by solving a static game. # Manufacturing-Marketing Interface Session: MA04 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: MSOM Track: Cluster: Room: Chair: Alysse R. Morton Chair Address: University of Utah, Eccles Sch. of Bus., 1645 East Campus Center Dr., Salt Lake City, UT 84112-9304 Chair E-mail: mgtarm@business.utah.edu Chair: Chair Address: Chair E-mail: MA04.1 Quality & Time-to-Market Trade-Offs when There are Multiple Product Generations • Leslie Olin Morgan; University of Utah, Eccles Sch. of Bus., 1645 East Campus Center Dr., Salt Lake City, UT 84112-9403; mgtlm@business.utah.edu • Ruskin Morgan; University of Utah, Eccles Sch. of Bus., 1645 East Campus Center Dr., Salt Lake City, UT 84112-9403; mktrm@business.utah.edu • William L. Moore; University of Utah, Eccles Sch. of Bus., 1645 East Campus Center Dr., Salt Lake City, UT 84112-9403; mktbm@business.utah.edu This research considers the quality versus time-to-market trade-off in the context of multiple product generations, extending previous work that focused on a single new product generation. We examine the factors affecting optimal time-to-market, and note differences from findings generated in research on the single-generation problem. MA04.2 The Impact of Design & Process Capabilities on Time-to-Market Performance • Janice Carrillo; Washington University, Olin Sch. of Bus., CB 1133, 1 Brookings Dr., St. Louis, MO 63130-4899; carrillo@suolin.wustl.edu • Richard M. Franza; Bentley College, Dept. of Mgmt., 175 Forest St., Waltham, MA 02452-4705; rfranza@bentley.edu Firms must develop appropriate design and production capabilities to bring valuable new products to market in an efficient manner. We present a model that examines the relationship between a firm's design and production capabilities and its time-to-market performance. Specifically, we analyze factors influencing a firm's ramp-up time and ultimate profitability. MA04.3 Product Line Expansion, Supply Chain Choice & Firm Survival • Taylor R. Randall; University of Utah, Eccles Sch. of Bus., 1645 East Campus Center Dr., Salt Lake City, UT 84112-9403; acttr@business.utah.edu We examine the path-dependent effect of sourcing decisions on the strategies companies use to expand their product lines. Empirical evidence which links coherent decisions about supply chain choice and product line strategy to firm survival is presented. # Probabilistic Models in Networks Session: MA05 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: Applied Probability Society Track: Cluster: Room: Chair: Harold Mortazavian Chair Address: UCLA, Computer Sci. Dept., Los Angeles, CA Chair E-mail: mor@cs.ucla.edu Chair: Chair Address: Chair E-mail: MA05.1 Traffic & Mobility Modeling in Wireless Networks • Brian Mark; George Mason University, Dept. of Elect. & Comp. Eng., 4400 University Dr., Fairfax, VA 22030-4444; bmark@gmu.edu • Shun-Zheng Yu; George Mason University, Dept. of Elect. & Comp. Eng., 4400 University Dr., Fairfax, VA 22030-4444; • Hisashi Kobayashi; Princeton University, Dept. of Elect. Eng. & CS, Princeton, NJ 08544-5263; hisashi@ee.princeton.edu Most previous work on traffic modeling and service provisioning has not thoroughly considered the impact of mobility in wireless networks. We develop an integrated set of probabilistic models for traffic and mobility in wireless networks. We discuss applications of these models to quality-of-service provisioning and fast web access at the wireless/wired network interface. MA05.2 A Unified Approach to the Analysis of Teletraffic Models & their Computational Solution • Khosrow Sohraby; University of Missouri, Comp. Sci. Telecom. Dept., 5100 Rockhill Rd., Kansas City, MO 64110-2499; sohraby@cstp.umkc.edu We present a system-theoretic approach to the solution of a large class of tele-traffic problems. Under the assumption of rationality of the generating functions of input processes and having obtained a state-space realization of the system, we derive an efficient matrix-geometric solution for the state probability vector of the system. MA05.3 An Approach to Distributed Probabilistic Routing & Congestion Control in Networks • Harold Mortazavian; UCLA, Computer Sci. Dept., Los Angeles, CA; mor@cs.ucla.edu We present a new approach to distributed probabilistic routing and congestion control in networks in which routers are viewed as local controllers. The concepts of probabilistic controllability and observability with specified parameters and a differential equation, depending on these parameters, whose solution provides that the probability distribution of routing is introduced at each node. MA05.4 Continuous Flow Models: Modeling, Simulation, IPA & Telecommunications Approaches • Benjamin Melamed; Rutgers University, Faculty of Mgmt., 265 Levin Bldg. Rockafeller Rd, New Brunswick, NJ 08903; melamed@rutcor.rutgers.edu We introduce a simple class of fluid-flow models, CFMs, reporting on joint work with Yorai Wardi. We will discuss problems of modeling propagation delays, loss-related performance measures, real-time computation and other pertinent issues. Also, the relation with infinitesimal perturbation analysis will be discussed. # Transit Applications of Operations Research Session: MA06 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: Transportation Science Section Track: Cluster: Room: Chair: Mark Hickman Chair Address: University of Arizona, Civil Eng. & Eng. Mechanics, PO Box 210072, Tucson, AZ 85721-0072 Chair E-mail: mhickman@engr.arizona.edu Chair: Chair Address: Chair E-mail: MA06.1 Hybrid Paratransit Delivery Methods • Majid Aldaihani; University of Southern California, Dept. of ISE, Los Angeles, CA 90089-0193; • Maged Dessouky; University of Southern California, Dept. of ISE, University Park Campus, Los Angeles, CA 90089-0193; maged@rcf.usc.edu • Randolph W. Hall; University of Southern California, Dept. of ISE, 3715 McClintock Ave., Los Angeles, CA 90089-0193; rwhall@mizar.usc.edu There is renewed interest in developing efficient paratransit methods for the elderly and disabled. Due to their convenience, most passengers prefer an on-demand service delivery method. This type of service may not be the most cost-effective method of transporting people. We present a hybrid delivery method consisting of both on-demand and fixed route services... MA06.2 Optimal Dispatching Strategies for Demand Responsive Systems • Guntram Noeth; University of Southern California, Bremenweg, Wuerzburg, Barvaria, 97084 , Germany; guntram_noeth@hotmail.com • Maged Dessouky; University of Southern California, Dept. of ISE, University Park Campus, Los Angeles, CA 90089-0193; maged@rcf.usc.edu • Randolph W. Hall; University of Southern California, Dept. of ISE, 3715 McClintock Ave., Los Angeles, CA 90089-0193; rwhall@mizar.usc.edu New communication technologies allow transit agencies to dispatch new requests in real time. For simple environments, we present some rules how transit systems are dispatched optimally to serve incoming customers. We also show how these strategies could be applied to more complex real-world environments. MA06.3 Rethinking Holding as a Transit Operations Control Strategy • Mark Hickman; University of Arizona, Civil Eng. & Eng. Mechanics, PO Box 210072, Tucson, AZ 85721-0072; mhickman@engr.arizona.edu Many public transit agencies use holding as a real-time control strategy to improve operations. Past studies have shown that these holding decisions are very sensitive to several critical assumptions about transit service. Using a broad sensitivity analysis, we examine the conditions under which holding is and is not advantageous. # Models For & About Education Session: MA07 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Mona Whitley Howard Chair Address: Paul Quinn College, 3837 Simpson Stuart Rd., Dallas, TX 75241 Chair E-mail: pqcmhoward@hotmail.com Chair: Chair Address: Chair E-mail: MA07.1 Resource Allocation in Academia Resource allocation issues arise frequently in academic settings. Examples include recruiting, salary adjustments and research proposal funding. The problems, as well as the solution approach, vary from situation to situation. The objective, however, is generally to find solutions that are fair. We focus on stronger formulations and more efficient solutions. MA07.2 Logarithmic Transformations in Regression: Some Issues You decide to use logarithmic transformations on the variables of a regression model. The scatter after transforming looks linear, r-squared has increased, the t-test for the slope has improved;therefore, you expect narrower prediction intervals for the same level of confidence. Well ... not necessarily. We also discuss other related issues. MA07.3 Process Simulation in Excel with SimQuick We introduce SimQuick, an Excel add-in that allows the user to simulate a variety of simple processes, e.g., queues, inventory, manufacturing, projects, etc. It's designed to be easy to learn for students in classes not devoted to simulation, e.g., operations management, MS, supply chain management, etc. MA07.4 Academic Decision Making using Two Forms of Conjoint Analysis Six attributes of service for non-traditional academic programs are evaluated using focus studies and two different forms of cohort analysis. Despite somewhat divergent outcomes, the results provide meaningful, customer-centered information that is useful for executive-level decision makers to improve services at an educational institution. # Panel: E-Business Strategy & Technology Developments in Practice Session: MA08 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: E-Business Section Track: Cluster: Room: Chair: Christopher L. Huntley Chair Address: Fairfield University, Sch. of Bus., Fairfield, CT 06430 Chair E-mail: chuntley@fair1.fairfield.edu Chair: Aaron Seymour Chair Address: Fairfield University, Sch. of Bus., Fairfield, CT 06430 Chair E-mail: seymour@fair1.fairfield.edu MA08.1 Panel: E-Business Strategy & Technology Developments in Practice Following a presentation on e-business strategy and technological development theory, our distinguished panel of e-business executives will address their respective view on e-business in practice. They will discuss their online firm's evolutionary process required to maintain competitiveness in their respective markets. In addition, they will speak to current opportunities and challenges faced by online firms in the Connecticut Corridor. # Uncertainty & Forecasting Techniques in Revenue Management Session: MA09 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: Revenue Management Section Track: Cluster: Room: Chair: Ahmet Kuyumcu Chair Address: Talus Solutions, Inc., 4751 Best Rd., Waterstone, Ste. 300, Atlanta, GA 30337-5609 Chair E-mail: akuyumcu@talussolutions.com Chair: Chair Address: Chair E-mail: MA09.1 Winning in a Competitive Bidding Environment • Thomas Qi; Talus Solutions, Inc., 4751 Best Rd., Waterstone, Ste. 300, Atlanta, GA 30337; tqi@talussolutions.com Uncertainty is the nature of any bidding games. Understanding yourself and your competitors is the key to winning in a competitive bidding environment. We discuss a multi-regression approach that forecasts win probabilities based on forecasts of own and competitor prices MA09.2 Forecasting Demand for the Natural Gas Pipeline Industry • Doug Harvey; PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006; dharvey@prosrm.com • Nicola Secomandi; PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006; nseconamdi@prosrm.com We will discuss the problem formulation and techniques of forecasting demand for firm transportation contracts on natural gas pipelines. We especially focus on mapping the demand for historical contracts to demand for a group of representative contracts expected to be in high demand in the future. MA09.3 A Gaming Twist in Hotel Revenue Management Similar RM concepts are applied differently across different industries. While leveraging much of the technology developed for the hotel industry, the extension of RM to manage room revenues for the gaming industry presents many new technical challenges. Some of these challenges, which include customer segmentation, special events and promotions, are explored. # Real-Life Scheduling Models Session: MA10 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Scheduling Room: Chair: Alessandro Agnetis Chair Address: Universita di Siena, Dipt. di Ingegneria & Info., via Roma 56, Siena, 53100 , Italy Chair E-mail: agnetis@dii.unisi.it,, http://www.dii.unisi.it/~agnetis Chair: Chair Address: Chair E-mail: MA10.1 Practical Scheduling Solutions • Anand Iyer; i2 Technologies, Inc., 909 East Las Colinas Blvd., Ste. 16, Irving, TX 75039; iyer@i2.com We will discuss the solution of scheduling problems that arise in the context of an integrated supply chain. The links between master planning, factory planning and scheduling will also be explored. MA10.2 Open Issues on Cyclic Scheduling • Walter Ukovich; Universita di Trieste, Via Valerio 10, Trieste, 34127 , Italy; • Piero Persi; Universita di Trieste, via Valerio 10, Trieste, 34127 , Italy; • Raffaele Pesenti; Universita di Palermo, Automatica e Sistemistica, viale delle science, Palermo, 90128 , Italy; Cyclic scheduling is an open field of research, involving different applications: manufacturing, robotic cells, flow lines, hoist scheduling, loop scheduling on parallel processors and timetabling. Besides providing an extensive survey, we analyze open issues not yet covered in the cyclic scheduling literature. MA10.3 Scheduling in Dual Gripper Robotic Cells for Productivity Gains • Chelliah Sriskandarajah; University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO4.7, Richardson, TX 75083-0688; • Suresh P. Sethi; University of Texas at Dallas, Sch. of Mgmt., Box 830688, Richardson, TX 75083-0688; • Jeffrey B. Sidney; University of Ottawa, Faculty of Admin., Ottawa, Ontario, K1N 6N5 , Canada; We study the problem of scheduling robot moves in dual gripper robot cells in a bufferless environment. The objective is steady-state throughput maximization. We show that an m-machine dual gripper robot cell may have at most twice the productivity of its single gripper counterpart. MA10.4 Resource Scheduling by Column Generation: A Real-Life Application • Claudio Arbib; Universita degli Studi di L'Aquila, Dipt. Matematica Pure e Appl., via Vetoio, Coppito, 67010 , Italy; • Gianfranco Ciaschetti; Universita degli Studi di L'Aquila, Dipt. Matematica Pure e Appl., via Vetoio, Coppito, 67010 , Italy; • Fabrizio Rossi; Universita degli Studi di L'Aquila, Dipt. Matematica Pure e Appl., via Vetoio, Coppito, 67010 , Italy; rossi@univaq.it • Stefano Smriglio; Universita degli Studi di L'Aquila, Dipt. Matematica Pure e Appl., via Vetoio, Coppito, 67010 , Italy; A manufacturing plant employs a number of parallel workstations to produce semifinite plastic parts by moulding. We analyse the problem of finding a schedule of workers and orders under several optimization criteria, propose a solution approach based on column generation and describe a simulation study. # Algorithm Design for Discrete Optimization Problems Session: MA11 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Integer Programming Room: Chair: Sheldon H. Jacobson Chair Address: University of Illinois, 1206 West Green St., MEB 226, MC 244, Urbana, IL 61801-2906 Chair E-mail: shj@uiuc.edu,, http://www.staff.uiuc.edu/~shj/shj.html Chair: Chair Address: Chair E-mail: MA11.1 Managerial Insights into the Team-Building Problem from a Combinatorial Optimization Model • Daniel Solow; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106; dxs8@po.cwru.edu • George L. Vairaktarakis; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106-7235; gxv5@po.cwru.edu • Sandy Kristin Piderit; Case Western Reserve University, Organizational Behavior Dept., WSOM, Cleveland, OH 44106; kep2@po.cwru.edu • Ming-Chi Tsai; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106; A combinatorial optimization model is presented for building a team in which interactions between individuals can be controlled. Certain versions are shown to be polynomially solvable while others are strongly NP-complete. Heuristics are developed to study the effect of replacing an individual on the team. Analysis and simulations provide managerial insights into factors that affect the average performance of the best team... MA11.2 A Modifid Benders' Partitioning Approach for Stochastic Programs with Mixed-Integer Recourse • Hanif D. Sherali; Virginia Polytechnic Institute & State University, Dept. of ISE, 0118, Blacksburg, VA 24061-0118; hanifs@vt.edu • Barbara M. P. Fraticelli; Virginia Polytechnic Institute & State University, Dept. of ISE, Mail Code 0118, Blacksburg, VA 24061; barbf@mail.vt.edu A modified Benders' decomposition approach is proposed for discrete optimization problems, including stochastic IPs, using the reformulation-linearization technique or lift-and-project cuts to construct a suitable partial convex hull representation as needed. The cutting planes are derived as functions of the first-stage variables to be globally valid and re-usable for all subproblems. MA11.3 Computational Analysis of a Flexible Assembly System Design Problem We present a heuristic, the pick and rule (PAR) heuristic, for designing flexible assembly systems. A lower bound algorithm and a direct B&B algorithm are formulated to assess the effectiveness of the PAR heuristic. Computational results with the PAR heuristic and these 2 algorithms are reported. # Purchasing & Supplier Selection Session: MA12 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Logistics & Supply Chain Management Room: Chair: Linda L. Stanley Chair Address: Our Lady of the Lake University, Sch. of Business, 411 Southwest 24th St., San Antonio, TX 78207 Chair E-mail: stanl@lake.ollusa.edu Chair: Chair Address: Chair E-mail: MA12.1 The Impact of NAFTA on Purchasing & the Supply Chain: A Longitudinal Study • Julie J. Gentry; University of Arkansas, Dept. of Mktg. & Transport, 302 BA; Academicians and industry experts originally forecast the NAFTA to dramatically alter the purchasing and supply chain management functions of US-based organizations. We will provide the research results from data collected in 2 surveys (1996 & 2000) investigating purchasing implications for US-based firms as a result of the NAFTA. MA12.2 A Review of the Product-Related Characteristics affecting a Manufacturing Firm's Decision to Select Source Suppliers • Mohammad Z. Bsat; Jackson State University, Dept. of Mgmt. & Mktg.; • Astrid M. Beckers; Middle Georgia College, 1100 Second St. SE, Cochran, GA 31014; We examine the topic of supplier selection from a product-related characteristics standpoint. The different characteristics that lend importance to the process of supplier selection are reviewed. This research emphasizes the importance of such characteristics in the supplier selection decision. MA12.3 An Examination of Supply Chain Management Practices in the Mexican Automotive Parts Industry: An Empirical Approach • Daniel A. Glaser-Segura; Our Lady of the Lake University, Sch. of Business, 411 SW 24th St., San Antonio, TX 78207; glasd@lake.ollusa.edu In a survey of Mexican automotive parts industry, managers reported the status of various supply chain management practices at their firms. The pattern of implementation is inconsistent with theoretical propositions established in industrialized nations and warrants review of practice and research in less developed countries. MA12.4 Relationship Management in the Supply Chain: A Comparative Analysis • Linda L. Stanley; Our Lady of the Lake University, Sch. of Business, 411 Southwest 24th St., San Antonio, TX 78207; stanl@lake.ollusa.edu We will provide a comparative analysis between the food processing and electronics industries regarding supplier management issues from a purchasing perspective. The research will focus on important criteria in a successful relationship and practices most commonly utilized today to manage the relationship. # Manufacturing Logistics Scheduling & Routing Session: MA13 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Manufacturing & Logistics Room: Chair: Zhi-Long Chen Chair Address: University of Pennsylvania, Dept. of Systems Eng., 220 South 33rd St., Philadelphia, PA 19104-6315 Chair E-mail: zlchen@seas.upenn.edu Chair: Chair Address: Chair E-mail: MA13.1 Magazine Configurations for a Single Machine with Finite Tool Capacity • Moshe Dror; University of Arizona, Coll. of Bus. & Public Admin., Dept. of MIS, Tucson, AZ 85721; mdror@mail.bpa.uarizona.edu • Andre Langevin; Ecole Polytechnique, Dept. of Math. & IE, CP 6079, Succ. Centre-ville, Montreal, Quebec, H3C 3A7 , Canada; andrel@crt.umontreal.ca • Francois Soumis; GERAD, Ecole Polytech., 3000 Cote-Sainte-Catherine, Montreal, Quebec, H3T 2A7 , Canada; soumis@crt.umontreal.ca • Pontien Mbaraga; Ecole Polytechnique de Montreal, GERAD, Montreal, Quebec, H3C 3A7 , Canada; We examine the problem of scheduling composite items on a single multipurpose machine with a finite capacity tool magazine. We present a novel shortest path with side constraints solution approach and compare it with solutions based on a TSP formulation. MA13.2 The Effectiveness of Zero Inventory Ordering Policies for a One-Warehouse Multi-Retailer Problem with Piecewise Linear Costs We consider a problem faced by companies that rely on truckload and carriers for the distribution of products across their supply chain. We analyze the effectiveness of the ZIO policies and develop approximation algorithms in this class. MA13.3 Dynamic Crane Deployment at Container Terminals • Raymond K. Cheung; HKUST, Dept. of IE/EM, Clear Water Bay, Kowloon, Hong Kong, , PR China; rcheung@ust.hk • Chung Lun Li; ; • Jeff Lin; ; We formulate the problem of deploying cranes among container blocks at a container terminal to minimize the workload unfinished on time as an integer program. We propose a relaxation-based method and a nonlinear approximation procedure to solve the problem. Experimental results will be discussed. MA13.4 Solving Practical Vehicle Routing Problems • Henry Xu; University of Pennsylvania, Dept. of Systems Eng., 220 South 33rd St., Philadelphia, PA 19104-6315; hxu@seas.upenn.edu • Zhi-Long Chen; University of Pennsylvania, Dept. of Systems Eng., 220 South 33rd St., Philadelphia, PA 19104-6315; zlchen@seas.upenn.edu • Srinivas Rajagopal; Manugistics, Inc., 585 East Swedesford Rd., Wayne, PA 19087; srajagop@manu.com • Sundar Arunapuram; Manugistics, Inc., 585 East Swedesford Rd., Wayne, PA 19087; sarunapu@manu.com We investigate real-world vehicle routing problems that involve complex cost structures and complex business constraints such as DOT rules, multiple time windows, nested precedence, and compatibility issues. We propose several optimization based algorithms to tackle such problems. # Airline Marketing & Planning Session: MA14 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: Aviation Applications Section Track: Cluster: Room: Chair: Milind G. Sohoni Chair Address: Delta Airlines, Atlanta, GA Chair E-mail: milind.sohoni@delta-air.com Chair: Chair Address: Chair E-mail: MA14.1 no show MA14.2 Revenue Management with Overbooking The bid-price revenue management model assumes as an input a set of adjusted flight leg capacities. These adjusted flight leg capacities are derived from physical available capacities by solving an overbooking optimization problem to handle cancellations and no-shows. We propose a bilevel optimization model for overbooking with the bid-price revenue management problem (a linear program) as a constraint. MA14.3 Bayesian Approaches to Forecasting & Seat Allocation • Chris K. Anderson; University of Western Ontario, Ivey Sch. of Bus., 1151 Richmond St., London, Ontario, N6A 3K7 , Canada; canderson@ivey.uwo.ca,, http://www.ivey.uwo.ca • Peter C. Bell; University of Western Ontario, Ivey Sch. of Bus., 1151 Richmond St., London, Ontario, N6A 3K7 , Canada; pbell@ivey.uwo.ca • John G. Wilson; University of Western Ontario, Ivey Sch. of Bus., 1151 Richmond St., London, Ontario, N6A 3K7 , Canada; A Bayesian model for the dynamic allocation of airline seats to various fare classes is developed. Seats sold, time to departure and historical information are all incorporated. The Bayesian framework allows for the explicit consideration of constrained demand and dependency across time periods and fare classes. MA14.4 Full vs. Partial Information about Customer Requests in Single-Leg Revenue Management • Darius Walczak; University of British Columbia, Operations & Logistics, 2053 Main Mall, Vancouver, BC, V6T 1Z2 , Canada; darius.walczak@commerce.ubc.ca • Shelby Brummelle; University of British Columbia, Operations & Logistics, 2053 Main Mall, Vancouver, BC, V6T 1Z2 , Canada; shelby.brumelle@ubc.ca A customer requests multiple seats at one of the fares, may not show up, can cancel and be overbooked. Customer fare and request size are not known but their joint probability distribution is available. We model this situation as a semi-Markov control problem with general state space and provide conditions for a structured optimal policy... # Service Parts Reengineering at IBM Session: MA15 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Quantitative Models in Supply Chain Management Room: Chair: George R. Wilson Chair Address: Lehigh University, IMSE Dept., Mohler Lab., 200 West Packer Ave., Bethlehem, PA 18015 Chair E-mail: grw3@lehigh.edu Chair: Chair Address: Chair E-mail: MA15.1 Customer Delivery Time Sensitive Inventory Network Design • Mark C. Booth; IBM, Service Parts Solutions, 5267 East Simpson Ferry Rd., Mechanicsburg, PA 17050; boothm@us.ibm.com • Erhan Kutanoglu; University of Arkansas, 4207 Bell Eng. Ctr., Dept. of IE, Fayetteville, AR 72701; erhank@engr.uark.edu • Terry Hammaker; ; IBM's service parts stocking policy has been based on fill rates but has migrated to on-time delivery to customers. A time-based service criterion will drive not only stocking levels but relative stocking locations, as well. The presentation emphasizes the impact of an on-time delivery policy on inventory network design. MA15.2 Inventory Neighborhoods: A Time Sensitive Service Parts Stocking Approach • George R. Wilson; Lehigh University, IMSE Dept., Mohler Lab., 200 West Packer Ave., Bethlehem, PA 18015; grw3@lehigh.edu • Selcuk Avci; Lehigh University, IMSE Dept., 200 Mohler Lab., 200 West Packer Ave., Bethelehem, PA 18015; • Francisco Barahona; IBM Watson Research Center, PO Box 218, Yorktown Heights, NY 10598; baranon@us.ibm.com • Yesim Erke; ; • Matthew Galati; ; • Christina Ma; ; IBM has migrated to a stocking policy that incorporates least cost lateral movement of parts from neighboring stocking points to meet time-based performance goals. Parts are clustered into serviceable machine groups across geographies and delivery within a parts grouping must be on time a prescribed proportion of the time. MA15.3 Dynamic Work-Force Scheduling • Brenda Dietrich; IBM TJ Watson Research Center, PO Box 218, Yorktown Heights, NY 10598; dietric@us.ibm.com • Baruch M. Schieber; IBM TJ Watson Research Center, PO Box 218, Yorktown Heights, NY 10598; sbar@watson.ibm.com IBM Global Services employs about 6,000 customer service representatives (CSRs) responding to more than 20,000 service calls per day. IBM developed an assignment engine that assigns calls to CSRs, striving to optimize several business objectives: maximize customer satisfaction, minimize CSR idle and travel times, minimize cost of parts shipment. # Applications of Multicriteria Tools in Single & Multiple Decision Maker Environments Session: MA16 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: MCDM Room: Chair: Peri H. Iz Chair Address: Health Care Finance Administration, Strategic Plng. & Evaluation, 7500 Security Blvd, MS C31926, Baltimore, MD 21244 Chair E-mail: piz@hcfa.gov,, izperi@iamdigex.net Chair: Chair Address: Chair E-mail: MA16.1 Employee Selection under Anti-Discrimination Law: Implications for Multicriteria Group Decision Support • Lorraine R. Gardiner; Auburn University, Dept. of Mgmt., Coll. of Bus., 415 West Magnolia Ave., Auburn, AL 36849; gardilr@auburn.edu • Debra Armstrong-Wright; Auburn University, Affimative Action/Equal Empl., 13 Quad Ctr., Auburn, AL 36849; armstde@auburn.edu Multicriteria group decision support offers methodology to committees involved in hiring decisions that can improve the chances for nondiscriminatory selection processes. The legal requirements for nondiscriminatory hiring decisions raise new challenges for both multicriteria decision aid methodology and the way it is applied in a group setting. MA16.2 Applying Multicriteria Analysis to the Upgrading of Instructional Computing Laboratories Universities are faced with periodically upgrading their instructional computing laboratories in response to increased user expectations and many enhanced hardware and software capabilities. Using student and faculty responses, multicriteria decision analysis can be used to assist in making more effective and acceptable choices from the wide variety of available alternatives. MA16.3 Benchmarking Supply Chain Management: The Use of Judgmental Models We examine the decision making process in supply chain management. We describe qualitative and quantitative approaches to support effective decision-making and explores metrics and mapping tools to compare supply chain performance. Some specific examples will be discussed to highlight critical issues # Tutorial: Integrated Supply Chain Planning for Semiconductor Industries Session: MA17 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Performance of Supply Networks Room: Chair: Mahesh Rajasekharan Chair Address: i2 Technologies, Inc., High Tech. Industry Bus. Unit Chair E-mail: Chair: Ben Mathias Chair Address: i2 Technologies, Inc., High Tech. Industry Bus. Unit Chair E-mail: MA17.1 Tutorial: Integrated Supply Chain Planning for Semiconductor Industries • Mahesh Rajasekharan; i2 Technologies, Inc., High Tech. Industry Bus. Unit; • Ben Mathias; i2 Technologies, Inc., High Tech. Industry Bus. Unit; Semiconductor companies are characterized by complex supply chains involving many different in-house fabrication and assembly facilities, foundries and subcontractors. The business environment is characterized by long manufacturing cycle times and short product life cycles. Hence, accurate forecasting, enterprise wide supply chain planning, strategic fulfillment and delayed product differentiation are critical for success... # E-Business: Strategy & Policy Session: MA18 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: E-Commerce Room: Chair: Sirkka Jarvenpaa Chair Address: University of Texas, Ctr. Bus., Technology & Law, MSIS Dept, CBA 5.202, Austin, TX 78712-1175 Chair E-mail: sjarvenpaa@mail.utexas.edu Chair: Chair Address: Chair E-mail: MA18.1 Open- vs. Closed-E-Business Models • Sirkka Jarvenpaa; University of Texas, Ctr. Bus., Technology & Law, MSIS Dept, CBA 5.202, Austin, TX 78712-1175; sjarvenpaa@mail.utexas.edu • Emerson Tiller; University of Texas, Ctr. Bus., Technology & Law, MSIS Dept., CBA 5.202, Austin, TX 78712; emerson.tiller@bus.utexas.edu The closed-Internet vs. open-Internet debate is not just a philosophical debate, but one that has profound ramifications to the viability of different business plans on the Internet. We provide a framework for managers to think about their business strategies in the light of this open- and closed-Internet debate. MA18.2 Internet Consumer Deception • Stefano Grazioli; University of Texas, Business Sch., Austin, TX 78712; • Sirkka Jarvenpaa; University of Texas, Ctr. Bus., Technology & Law, MSIS Dept, CBA 5.202, Austin, TX 78712-1175; sjarvenpaa@mail.utexas.edu Selling defective goods at an on-line auction and touting under-performing securities are specific examples of the broad phenomenon of Internet consumer deception. We report on the creation of a database of cases of Internet consumer deception as a first step towards a broader investigation of this form of business misconduct. # Innovative Management of Manufacturing Capacity Session: MA19 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Technology Innovations & Operations Room: Chair: James R. Bradley Chair Address: Cornell University, Johnson Grad. Sch. of Mgmt., 321 Sage Hall, Ithaca, NY 14853-6201 Chair E-mail: jrb28@cornell.edu Chair: Chair Address: Chair E-mail: MA19.1 Safety-Capacity Allocation in Lean Manufacturing • John J. Liu; University of Wisconsin, Sch. of Bus., Milwaukee, WI 53201; jjl@uwm.edu Lean manufacturing, one of the most fast-spreading manufacturing innovations, is centered on the elimination of wastes. A major source of wastes is in safety capacity, which refers to precautionary production capacity against uncertainty. How to allocate safety-capacity is of great importance and difficulty especially when the system is 'lean'... MA19.2 Increasing Product Mix Flexibility on the Assembly Line Assembly lines are typically managed with the goal of attaining high labor utilization, which makes the assembly line inflexible to product variety. We discuss how slack capacity can be used in order to increase product mix flexibility. The divergent approaches toward slack capacity in this paper and the previous paper demonstrate that 'different approaches are appropriate for different goals.' MA19.3 Managing Capacity when Demand Fluctuates in Mix & Volume: Field Data from Toyota Systems Managing Settings We present field-data that show that Toyota Production. System-managed organizations address fluctuations in demand and capacity, both mix and volume, in ways that are counter-intuitive. Inventory is not pooled, machines and people are dedicated to particular production pathways rather than letting jobs get done on a first-available basis and the span of managerial control is relatively narrow and steep, not broad and flat. # New Advances & Applications of Scatter & Tabu Search Session: MA20 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Tabu & Scatter Search Room: Chair: Manuel Laguna Chair Address: University of Colorado, School of Bus., CB 419, Boulder, CO 80309-0419 Chair E-mail: manuel.laguna@colorado.edu Chair: Chair Address: Chair E-mail: MA20.1 A Tabu Search Algorithm for Multi-Objective Optimization: An Application to Flowshop Scheduling • Vinicius A. Armentano; Universidade Estadual de Campinas, FEEC-UNICAMP, CP 6101, Campinas Sao Paulo, 13083-970 , Brazil; • Jose Elias C. Arroyo; Universidade Estadual de Campinas, FEEC-UNICAMP, CP 6101, Campinas Sao Paulo, 13083-970 , Brazil; We present a tabu search algorithm for finding a set of approximately nondominated solutions for multi-objective combinatorial optimization problems. The algorithm is applied to flowshop scheduling problems to minimize the makespan and the maximum tardiness. Its performance is compared with an exact method and a tabu search algorithm from the literature. MA20.2 Diversification with Multiple Neighborhoods • Ed Mooney; Montana State University, MIE Dept., Bozeman, MT 59717-3800; We will describe a search approach that simultaneously explores several neighborhoods at once and discuss the relationship to and integration with traditional tabu search methods. An example will be presented demonstrating application of the approach to a multiple resource scheduling problem. MA20.3 OQGRG: A Scatter Search Approach for Solving Constrained Non-Linear Global Optimization Problems • Zsolt Ugray; University of Texas, MSIS Dept., CBA 5.202, Austin, TX 78712-1175; • Leon S. Lasdon; University of Texas, MSIS Dept., CBA 5.202, Austin, TX 78712-1175; lasdon@mail.utexas.edu • John C. Plummer; University of Texas, MSIS Dept., CBA 5.202, Austin, TX 78712-1175; jcplummer@mail.utexas.edu We present a method of solving large-scale, constrained non-linear problems with integer variables. We show computational results of the algorithm that uses the combination of Optquest, a scatter search algorithm and LSGRG, a local solver. # Session V Session: MA21 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: SICUP Room: Chair: Jose Fernando Oliveira Chair Address: University of Porto, Dept. de Eng. Electro e Comp., Rua dos Bragas, Porto, 4050-123 , Portugal Chair E-mail: jfo@fe.up.pt,, jfo@inescn.pt Chair: Chair Address: Chair E-mail: MA21.1 Meta-Heuristics for Calculating 2- & 3-Dimensional Bin-Packing Problems • Elita A. Mukhacheva; Ufa State Aviation Technical University, Karl Marx Str. 12, Ufa, Bashkortostan, 450000 , Russia; elita@vmk.ugatu.ac.ru • Aida F. Valeeva; Ufa State Aviation Technical University, Karl Marx Str. 12, Ufa, Bashkortostan, 450000 , Russia; • Anna S. Mukhacheva; Ufa State Aviation Technical University, Karl Marx Str. 12, Ufa, Bashkortostan, 450000 , Russia; anna@vmk.ugatu.ac.ru Two- and 3-dimensional bin-packing problems are considered. These problems are known to belong to the class of NP-hard. GA and dynamic sorting stochastic algorithms are applied to solve 2- and 3-BPP. The results of numerical experiment on Waescher's method MA21.2 Genetic Algorithms for 2-Dimensional Packing & Floor Planning • Christine L. Valenzuela; Cardiff University, Dept. of Comp. Sci., PO Box 916, Cardiff, CF2 3XF , UK; christine@cs.cf.ac.uk • Pearl Y. Wang; George Mason University, Comp. Sci. Dept., MSN 4A5, 4400 University Dr., Fairfax, VA 22030-4444; pwang@cs.gmu.edu We present a GA that breeds normalized postfix expressions for packing rectangles. Experiments confirm that the approach is very effective and produces excellent results on a range of 2-dimensional packing and floor planning problems. Furthermore, the slicing construction process scales favorably when compared with existing procedures. MA21.3 Research of Heuristics Efficiency for 2-Dimensional Bin-Packing Problems • Anna S. Mukhacheva; Ufa State Aviation Technical University, Karl Marx Str. 12, Ufa, Bashkortostan, 450000 , Russia; anna@vmk.ugatu.ac.ru • Artem V. Chiglintsev; Ufa State Aviation Technical University, Karl Marx Str. 12, Ufa, Bashkortostan, 450000 , Russia; Problems of guillotine stock-cutting and rectangular packing in a semi-endless strip. Guillotine and block GAs are applied. In guillotine GAs, the genes are rectangles with suitable bin correlation strategy, the allels differ in block element rearrangement. The results of numerical experiment are presented. MA21.4 A Two-Exchange Nesting Heuristic for Nesting Problems • Jose Fernando Oliveira; University of Porto, Dept. de Eng. Electro e Comp., Rua dos Bragas, Porto, 4050-123 , Portugal; jfo@fe.up.pt,, jfo@inescn.pt • Antonio Miguel Gomes; University of Porto, Fac. of Eng., MSE Unit; We describe a new heuristic for the nesting problem based on a 2-exchange neighborhood generation strategy. This mechanism guides the search through the solution space, constituted by the sequences of pieces, relying on a low-level placement heuristic to actually convert one sequence in a feasible layout... # Clinical Applications of MS/OR Session: MA22 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: Health Applications Section Track: Cluster: Room: Chair: Stefanos Zenios Chair Address: Stanford University, Grad. Sch. of Business, Stanford, CA 94305-5015 Chair E-mail: stefzen@leland.stanford.edu Chair: Chair Address: Chair E-mail: MA22.1 The Use of Virtual Worlds & Animated Personas to Improve Health Care Knowledge & Self-Care Behavior: The HEART-SENSE Game • Barry G. Silverman; University of Pennsylvania, Systems Eng. Dept., 229C Towne Bldg., Philadelphia, PA 19104-6315; barryg@seas.upenn.edu • John Holmes; University of Pennsylvania, Clinical Epidemiology/Biostats, c/o 229c Towne Bldg., Philadelphia, PA 19104-6315; jholmes@cceb.med.upenn.edu Our goal is to determine whether a computer based training game, HEART-SENSE, can improve recognition of heart attack symptoms and shift behavioral issues so as to reduce pre-hospitalization delay in seeking treatment thereby reduce myocardial infarction mortality and morbidity. Innovations are: a design for a generic simulator package for promoting health behavior shifts and equations for animated pedagogical agents... MA22.2 Designing Radiotherapy Plans with Elastic Constraints & Interior Point Methods • Allen Holder; Trinity University, Math Dept., 715 Stadium Dr., San Antonio, TX 78212; aholder@trinity.edu A new LP model used to aid in the design of radiotherapy plans is introduced. This model incorporates elastic constraints, and when solved with a path following interior point method, produces favorable plans. A sound mathematical analysis shows how to interpret the solution; hence, the treatment planner receives meaningful knowledge about the radiotherapy plan being developed. Preliminary experiments are conducted. MA22.3 Nonrandom Mixing in Network Epidemic Models • Gregory S. Zaric; University of Western Ontario, Ivey Sch. of Business, London, Ontario, N6A 3K7 , Canada; gzaric@ivey.uwo.ca A common assumption of epidemic models is that mixing occurs randomly among all members of the population. Observations of networks of IDUs reveal that this is not always the case. We use a network epidemic model to compare random mixing with mixing models that account for some social structure. MA22.4 Outcomes-Adjusted Reimbursement in a Health Care Delivery System We consider a health care delivery system with 2 non-cooperative parties: a purchaser of medical services and a specialized provider. We develop a dynamic principal-agent model that captures the interaction between the 2 parties. The optimal payment system is identified and an application in Medicare's ESRD program is discussed. # Simulation Input & Output Analysis Session: MA23 Date/Time: Monday 08:15-09:45 Type: Invite/Sponsor Sponsor: Simulation Section Track: Cluster: Simulation Room: Chair: James R. Wilson Chair Address: North Carolina State University, Dept. of IE, Raleigh, NC 27695-7906 Chair E-mail: Chair: Chair Address: Chair E-mail: MA23.1 Chessboard Distributions & Random Vectors with Prescribed Marginals & a Covariance Matrix • Shane Henderson; University of Michigan, IOE Dept., 1205 Beal Ave., Ann Arbor, MI 48109-2117; shane.henderson@umich.edu • Soumyadip Ghosh; ; A covariance matrix is feasible for a given set of marginals if a random vector exists with the given marginal distributions. Using so-called chessboard distributions, we show that there are feasible covariance matrices for uniform marginals that cannot be matched using the NORTA method. A modification of NORTA is given for handling such a situation. MA23.2 A Discussion of Multivariate Ranking & Selection Procedures We begin with a discussion of the methods for multivariate ranking and selection found in the literature. We compare these methods with our current research that combines multiple attribute utility theory with scalar ranking and selection procedures. Finally, we discuss some future research directions. MA23.3 Improved Batching for Confidence Interval Construction in Steady-State Simulation • Natalie Steiger; University of North Carolina, Greensboro, NC; • James R. Wilson; North Carolina State University, Dept. of IE, Raleigh, NC 27695-7906; We describe the automated simulation analysis procedure (ASAP), which generates confidence intervals for steady-state simulation that satisfy a user-specified absolute or relative precision requirement. An experimental performance evaluation demonstrates the advantages of ASAP vs. other widely used batch-means procedures. MA23.4 Excel-Based Macros for Teaching Output Analysis Techniques & Concepts • M. C. Court; University of Oklahoma, Sch. of IE, 202 West Boyd, Rm. 124, Norman, OK 73072; mcourt@ou.edu Macros were developed for performing parameter estimation of average queue time and system time for M/M/1 queuing systems. The macros were deployed in 2 undergraduate IE courses. The result in the undergraduate simulation course was an improvement in student grades on assignments and test questions concerning simulation output analysis. # Warehousing & Distribution Session: MA24 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Warehousing & Distribution Room: Chair: Kevin R. Gue Chair Address: Naval Postgraduate School, Dept. of Systems Mgmt., Monterey, CA 93943 Chair E-mail: Chair: Chair Address: Chair E-mail: MA24.1 Facility Shapes for Crossdocks • Kevin R. Gue; Naval Postgraduate School, Dept. of Systems Mgmt., Monterey, CA 93943; • John Bartholdi, III; Georgia Institute of Techology, Sch. of ISyE, Atlanta, GA 30332-0205; john.bartholdi@isye.gatech.edu In addition to the popular rectangular shape for crossdocks, we have seen docks in the shape of an L, T, H and U. We show which shapes are best for different sizes of crossdocks and argue that for large docks, the best shape is none of the above. MA24.2 Container Sequencing to Facilitate Downstream Effectiveness • Michael J. Racer; University of Memphis, Dept. of ISE, Memphis, TN 38152; mracer@memphis.edu • Syed Usman; University of Memphis, Dept. of ISE, Memphis, TN 38152; usyed@memphis.edu We review the development of a technique for sequencing containers for offload in an express mail environment. Because the destination mix of items in containers is variable, the ordering of containers could affect temporal system performance downstream. We consider a technique for sequencing and evaluating conditions tools. MA24.3 An Integrated Warehouse Management System at YCH Logistics, Singapore YCH logistics, Singapore, manages a group of central distribution centers with storage systems tailored for each of its 15 clients operating in the South East Asia Pacific region. We discuss a warehouse management system that connects all the players in the supply chain via an EDI network. MA24.4 Keys to Successful Implementation of a Warehouse Management System • Lon Cross; Manhattan Associates, 2300 Windy Ridge Rd., 7th Floor, Atlanta, GA 30339; lcross@manh.com We present several examples of both successful and unsuccessful implementations and discuss key principles for success. In addition, we will describe our new 'e-fulfillment' methodology which allows us to fast-track implementations for e-commerce clients. # On the Efficient Solution of Mathematical Programming Problem Classes Session: MA25 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: Optimization Section Track: Cluster: Linear Programming & Complementarity Room: Chair: Georgia Perakis Chair Address: MIT, Sloan Sch. of Bus., OR Ctr., 50 Memorial Dr., Rm. E53-359, Cambridge, MA 02139 Chair E-mail: georgiap@mit.edu Chair: Chair Address: Chair E-mail: MA25.1 An Implementable Trust Region Method for Bi-Level Programming We consider the approximation of mathematical programs with variational inequality constraints by solvable programs of the same type, viz. bilevel programs involving linear approximations of the upper level objective and the lower level mapping. Based on these approximations, a trust region approach with convergence towards a 'strong' stationary point is presented. MA25.2 Solving Nonconvex Programs by the Successive Convex Relation Method The SCRM proposed by Kojima & Tuncel (1998) gives theoretical algorithms to solve nonconvex (quadratic) programs through successive LP, SDP or convex program relaxations. We propose some heuristic procedures for this method which work positively in practice for small-sized NCPs. MA25.3 Cutting Plane & Interior Point Methods for Solving Variational Inequalities • Marina Zaretsky; MIT, OR Ctr., 1 Amherst St., Rm. E40-130, Cambridge, MA 02139; mzar@mit.edu • Georgia Perakis; MIT, Sloan Sch. of Bus., OR Ctr., 50 Memorial Dr., Rm. E53-359, Cambridge, MA 02139; georgiap@mit.edu We considers methods for solving variational inequalities efficiently. Our approach incorporates cuts as well as long- and short-step line searches and as a result, allows us to solve a larger class of problems. Preliminary computational results are also encouraging. # Stochastic Models in Supply Chain Applications Session: MA26 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Stochastic Models & Applications Room: Chair: Jeannette Song Chair Address: University of California, Grad. Sch. of Mgmt., Irvine, CA 92697 Chair E-mail: jssong@uci.edu Chair: Chair Address: Chair E-mail: MA26.1 no show • Hari S. Abhyankar; PricewaterhouseCoopers, 144 Middlesex Turnpike, Burlington, MA 01803; hsabhyan@aol.com • Stephen C. Graves; MIT, 77 Massachusetts Ave., Rm. E40-439, Cambridge, MA 02139-4307; sgraves@mit.edu MA26.2 Optimal Replenishment Policies for Multi-Echelon Inventory Problems under Advance Demand Information • Ozalp Ozer; Columbia University, 500 West 120th St., MC 4704, New York, NY 10027-6699; ozalp@ieor.columbia.edu • Guillermo Gallego; Columbia University, IEOR Dept., 324 SW Mudd Bldg., 500 West 120th St., Rm. 331, New York, NY 10027; ggallego@ieor.columbia.edu Customers and downstream supply chain partners often place, or can be induced to place, orders in advance of future requirements. We show how to optimally incorporate advance demand information into periodic review multi-stage inventory systems in series. We also show that optimal policies are easy to compute when costs and demands are stationary. MA26.3 A Markovian Dual-Source Production-Inventory Model with Order Bands Motivated by applications in manufacturing and blood-supply industries, we generalize and unify various inventory models. Our model includes dual suppliers with differing lead-times, as well as state-dependent order size prescriptions, unit costs, inventory/penalty costs and demand distributions. We provide optimal policies under certain conditions; otherwise we validate heuristic methods. MA26.4 A Multi-Product Assemble-to-Order System with Random Lead Times: Performance Analysis, Approximations & Optimization • Yingdong Lu; University of California, Grad. Sch. of Mgmt., Irvine, CA 92697; lyd@gsm.uci.edu • Jeannette Song; University of California, Grad. Sch. of Mgmt., Irvine, CA 92697; jssong@uci.edu • David D. Yao; Columbia University, IEOR Dept., New York, NY 10027-6699; yao@ieor.columbia.edu We study a mulit-product assemble-to-order system. Customer demand for each product follows an independent Poisson process and replenishment lead times for each component are i.i.d. random variables. For any given base-stock policy, we conduct performance analysis to gain insights. We also develop easy-to-compute performance bounds and approximations and study optimization problems. # Project Management I Session: MA27 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Ted D. Klastorin Chair Address: University of Washington, Dept. of MS, Box 353200, Seattle, WA 98195-3200 Chair E-mail: tedk@u.washington.edu Chair: Chair Address: Chair E-mail: MA27.1 withdrawn - author request of 9/22 MA27.2 A Flexible Model for the Time/Cost Trade-Off Problem • John Moussourakis; Rider University, 2083 Lawrenceville Rd., Lawrenceville, NJ 08648; moussourakis@rider.edu • Cengiz Haksever; Rider University, 2083 Lawrenceville Rd., Lawrenceville, NJ 08648; haksever@rider.edu We present a flexible MIP model for the solution of the time/cost trade-off problem encountered in project management. Whereas it is commonly assumed that the time/cost function is linear, the model presented makes minimal assumptions and accommodates any type of cost function that is linear, piecewise linear or discrete. MA27.3 no show • Gunduz Ulusoy; Sabanci University, Sch. of Eng. & Natural Sci., Orhanli, Tuzla, Istanbul, 81474 , Turkey; gunduz@sabanciuniv.edu • Funda Serifoglu; Izzet Baysal University, Bolu, , Turkey; • Sule Sahin; Logo Software Development Co., Istanbul, , Turkey; MA27.4 Project Planning under the Threat of a Potential Catastrophe • Ted D. Klastorin; University of Washington, Dept. of MS, Box 353200, Seattle, WA 98195-3200; tedk@u.washington.edu • Prabhu K. Aggarwal; College of William & Mary, Sch. of Business, Williamsburg, VA 23185; pkagga@business.wm.edu We consider a project planning problem when there is a threat of a catastrophic event, e.g., strike, occurring sometime during the project. Should the project manager try to build excess slack into the project (in case the event occurs) or wait to see if the event actually occurs? We develop a model to analyze this situation and demonstrate some non-intuitive results. # Freight Applications Session: MA28 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: RASIG/Optimization Section Track: Network Flows Cluster: Room: Chair: Dharma R. Acharya Chair Address: CSX Transportation, 500 Water St., J-300, Jacksonville, FL 32082 Chair E-mail: dharma_acharya@csx.com Chair: Chair Address: Chair E-mail: MA28.1 A Network Model for Crew Scheduling within Railroads Currently, railroad crews do not get any scheduled time off. They can be called to work at any time as long as they are rested and on the active board. We present a model that can be used to analyze the operating effects of providing crews with structured time off. MA28.2 Application of Network Flow Models in Traffic Data Estimation & Passenger Rail Revenue Management • Marc S. Meketon; MultiModal Applied Systems, Inc., 125 Village Blvd., Ste. 270, Princeton, NJ 08540; marc@multimodalinc.com We demonstrate estimation of coal traffic flows from high quality mine and customer data and low quality OD flow data using matrix balancing and network flow algorithms. We demonstrate a network flow model to estimate 'bid prices' in a passenger rail revenue management and compare it to more complex models. MA28.3 no show • Yen S. Shan; Perseco, 3075 Highland Parkway, Downers Grove, IL 60515; yshan@perseco.com MA28.4 Applying the Inverse Optimization Technique to Railway Traffic Flow Networks We describe the application of the inverse optimization technique to optimally adjust link cost factors to closely match the historical multi-commodity traffic flow. When using various tools to simulate railway traffic flow, network link cost factors (impedances) are introduced to reflect preferred routes, which are often different from shortest paths. # Market Mechanisms for Competitive Electricity Session: MA29 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: ENRE Track: Cluster: Room: Chair: Hung Po Chao Chair Address: Electric Power Research Institute, 3412 Hillview Ave., PO Box 10412, Palo Alto, CA 94303 Chair E-mail: hchao@epri.com Chair: Chair Address: Chair E-mail: MA29.1 Pricing Ramping & Spinning Reserves in Competitive Electricity Markets • Ross Baldick; University of Texas, Dept. of Elect. & Comp. Eng., Austin, TX 78712-1084; baldick@ece.utexas.edu We formulate the pool problem of dispatching energy and spinning reserves. Unlike previous formulations, we also consider a bid price for ramping contribution, together with a maximum ramping capacity. Explicit pricing for 'ramping contribution' by a generator provides a bridging between the time-scales of hourly energy and the time-scales of ancillary services. We explore the implications for energy and reserve markets. MA29.2 The Next Generation of Unit Commitment Models • Benjamin F. Hobbs; JHU, 313 Ames Hall, DOGEE, 3400 North Charles St., Baltimore, MD 21218; bhobbs@jhu.edu • Michael H. Rothkopf; Rutgers University, RUTCOR, 640 Bartholomew Rd., Piscataway, NJ 08854-8003; rothkopf@rutcor.rutgers.edu • Richard P. O'Neill; Federal Energy Regulatory Commission, Office of Economic Policy, 888 First St. NE, Washington, DC 20426; richard.oneill@ferc.fed.us • Hung Po Chao; Electric Power Research Institute, 3412 Hillview Ave., PO Box 10412, Palo Alto, CA 94303; hchao@epri.com An EPRI-NSF workshop recently addressed the needs for improved formulations and solution algorithms for the unit commitment problem in a competitive environment. Recent developments concerning computational capabilities and roles for unit commitment models are summarized, as are the participants' recommendations for future R&D. MA29.3 Flow-Based Transmission Rights & Congestion Mangaement in Competitive Electricity Markets • Hung Po Chao; Electric Power Research Institute, 3412 Hillview Ave., PO Box 10412, Palo Alto, CA 94303; hchao@epri.com • Stephen C. Peck; Electric Power Research Institute, 3412 Hillview Ave., PO Box 10412, Palo Alto, CA 94303; • Shmuel S. Oren; University of California, Dept. of IE&OR, 4135 Etcheverry Hall, Berkeley, CA 94720; oren@ieor.berkeley.edu • Robert B. Wilson; Stanford University, Dept. of IEOR, Stanford, CA 94305; rwilson@stanford.edu We describe a design for electricity markets and RTOs based on a flow-based paradigm that unbundles energy and transmission service. Tradable rights to the transmission grid are characterized in terms of financial rights with scheduling priority to power flow on congestion prone flowgates. We discuss the advantages of this approach and implementation issues. MA29.4 Value of Reserves in Electricity Market Design • Hung Po Chao; Electric Power Research Institute, 3412 Hillview Ave., PO Box 10412, Palo Alto, CA 94303; hchao@epri.com In many reserve markets, reserves are not properly priced in the current single-settlement system resulting in capacity shortages, gaming and subsequent price spikes. I sketch an incentive-compatible approach to forward reserve pricing that incorporates the reliability-enhancing value of reserves. This design can be readily implemented in NEPOOL and is consistent with the market design principles for reserve markets in a multiple settlement system. # Tutorial: INFORMS Computing Society Spring 2000 Prize Session: MA30 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: Computing Society Track: Cluster: Room: Chair: Richard S. Barr Chair Address: SMU, Science Info. Ctr. 306, Dallas, TX 75275-0122 Chair E-mail: barr@seas.smu.edu Chair: Chair Address: Chair E-mail: MA30.1 Tutorial: Global Optimization in Action • Janos D. Pinter; Pinter Consulting Services/Dalhousie University, 129 Glenforest Dr., Halifax, Nova Scotia, B3M 1J2 , Canada; jdpinter@is.dal.ca Continuous GO is aimed at finding the global solution in general nonlinear models. We highlight the most important GO models, discuss adaptive partition strategies to solve (Lipschitz-) continuous GO problems, present the essential convergence results and key algorithm implementation aspects, describe the LGO software and review several interesting GO applications. # Computer Science & Internet Applications Session: MA31 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: John S. Hollywood Chair Address: RAND Corporation, 1200 South Hayes St., Arlington, VA 22202 Chair E-mail: johnsh@rand.org Chair: Chair Address: Chair E-mail: MA31.1 Visualizing Data with Adjacency Measures • S. Raghavan; University of Maryland, Smith Sch. of Bus., College Park, MD 20742-1815; raghavan@wam.umd.edu • Edward Condon; University of Maryland, Inst. for Plasma Research, College Park, MD 20742; econdon@glue.umd.edu • Bruce L. Golden; University of Maryland, Smith Sch. of Bus., College Park, MD 20742-1815; bgolden@uhsmith.umd.edu • Shreevardhan Lele; University of Maryland, Smith Sch. of Bus., College Park, MD 20742-1815; slele@rhsmith.umd.edu • Edward A. Wasil, Jr.; American University, Kogod Sch. of Bus., 4400 Massachusetts Ave. NW, Washington, DC 20016; ewasil@american.edu Adjacency measures are frequently available for data. For example, retailers track sets of the most similar (adjacent) items sold, e.g., CDs, books, etc. We discuss a simple technique to obtain distance measures between data points and thus visualize them using adjacency measures. An application to undergraduate colleges using Fiske's guide is discussed. MA31.2 A Model of Small Business Web Use & Benefits • Masoud Yasai-Ardekani; University of Wisconsin, Sch. of Bus. Admin., PO Box 742, Milwaukee, WI 53201-0742; yasai@uwm.edu • Ram Ramamurthy; University of Wisconsin; • Fatemeh Mariam Zahedi; University of Wisconsin, School of Bus. Admin., PO Box 742, Milwaukee, WI 53201; • Kurt Pflughoeft; Market Probe, Inc., Milwaukee, WI; Our model describes the relationships between organizational context, IT infrastructure, web use and web benefits for small businesses. Results indicate that market pressure and scope ofoperations are related to web use and its associated benefits through a responsive IT infrastructure. Implications for managers and web providers are discussed. MA31.3 Scheduling Issues in Layer 3 Protocol Processing • Soo Y. Chang; Pohang University of Science & Technology, Dept. of IE, Hyoja San 31, Pohang, Kyungbuk, 790-784 , South Korea; syc@postech.edu • Jaeyoung Choi; POSTECH, PIRL, Hyoja San 31, Pohang, Kyungbuk, 790-784 , South Korea; y2kchoi@postech.edu The layer 3 protocols are processed by the routers and layer 3 or higher switching equipment. We present the various scheduling issues that rise in the development of layer 3 protocol processing systems. We present our experience with the development of a managed switch with 24 fast ethernet and 2 giga ports. MA31.4 Scheduling Heuristics for Large, Heterogeneous Computing Networks • John S. Hollywood; RAND Corporation, 1200 South Hayes St., Arlington, VA 22202; johnsh@rand.org • Mary R. Eubanks; Booz-Allen & Hamilton, 134 National Business Parkway, Annapolis Junction, MD 20701; eubanks_mary@bah.com We present and provide simulation results for a set of scheduling heuristics tailored for large, heterogeneous computer networks with high job arrival rates (several per second). These heuristics combine elements of elementary computer science algorithms with more sophisticated heuristics seen in the scheduling literature. # Telecommunication Systems I Session: MA32 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Abdullah Konak Chair Address: Auburn University, Dept. of ISE, 207 Dunstan Hall, Auburn, AL 36849 Chair E-mail: akonak@eng.auburn.edu,, http://www.eng.auburn.edu/~akonak Chair: Chair Address: Chair E-mail: MA32.1 no show • Joseph B. Kroculick; Illinois Institute of Technology, 22 White Oak Dr., Jim Thorpe, PA 18229; krocjoe@acm.org • Cynthia S. Hood; Illinois Institute of Technology, Chicago, IL; MA32.2 A New Techno-Economic Modeling of High Speed Networks for the Internet We present a techno-economic analysis of high speed network technologies for the internet service. We develop a new modeling and analysis methodology to help network planner deploy high speed network. We present a preliminary result to show the robustness of the proposed modeling. MA32.3 Multiperiod Capacity Planning for the Telecommunication Network Design Problem • Ali Amiri; Oklahoma State University, 210 Coll. of Business, Stillwater, OK 74078; amiri@okstate.edu We study the multi-period capacity planning for the telecommunication network design problem that involves deciding where, when and how much transmission capacity needs to be installed, over a multi-period planning horizon, to meet increasing traffic requirements at minimum total discounted cost. A solution procedure is proposed and computational results are reported. MA32.4 A Multi-Objective Genetic Algorithm Approach for the Communication Network Design Problem Considering Reliability A multi-objective GA approach is proposed for the communication network design problem considering reliability. While GA guides the main search, heuristics and exact methods are employed to solve sub problems such as the capacity and route assignment problems. Efficient methods are used to evaluate network reliability. # Production & Scheduling I Session: MA33 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Viput Ongsakul Chair Address: Texas Technical University, 201 Indiana Ave., #321G, Lubbock, TX 79415 Chair E-mail: vongsaku@ttacs.ttu.edu Chair: Chair Address: Chair E-mail: MA33.1 Scheduling a Batch Processing Machine with Incompatible Job Families The problem of scheduling batch processors is important in some industries and, at a more fundamental level, captures an element of complexity common to many practical situations. We describe a B&B procedure applicable to a batch processor model with incompatible job families. The objective is to minimize total weighted flowtime... MA33.2 A Nonlinear Programming Formulation of Real-Time Scheduling Problems We examine a class of scheduling problems familiar in the design of real-time computer operating systems, but relatively unexplored in the OR literature. We formulate one of these problems as an NLP problem and show that its solution corresponds to the results given in the computer science literature. MA33.3 The Mixed Batch Scheduling Problem The problem of mixed batch scheduling arise from leather processing, where the special characteristic is that leathers are processed in a flow shop manner but consisting of a mix of batch and individual processes. Our objective is to find a schedule that minimizes makespan. # Decision Support Systems I Session: MA34 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Tanu S. Bhatnagar Chair Address: George Washington University, 403 Monroe Hall, 2115 G St. NW, Washington, DC 20052 Chair E-mail: tsb@gwu.edu Chair: Chair Address: Chair E-mail: MA34.1 withdrawn - author request of 11/4 • Kenneth A. Lenig; US Army, TRADOC Analysis Ctr., 14 Hancock Ave., Ft. Leavenworth, KS 66027; lenigk@trac.army.mil • John M. Harwig; US Army, 16602 Pocono Dr., Round Rock, TX; harwigj@utexas.edu MA34.2 GeoRedes: A Spatial Decision Support System for Vehicle Routing & Facility Location Problems GEOREDES is an SDSS that integrates GISs with techniques for solving VRPs and facility location problems. The system was built using the MapObjects library. Tabu search and simulated annealing heuristics are available for solving VRPs; an exact algorithm solves uncapacitated location problems. MA34.3 no show • Cemal Akyel; Middle East Technical University, Dept. of Bus. Admin., Inonu Bulvari, Ankara, 06531 , Turkey; akyel@ba.metu.edu.tr MA34.4 A Decision Support System for Fund Pricing Accountant Assignments A Boston-based mutual funds firm will soon be relocated to a facility with desk capacity substantially below the current staff level. A binary programming model will assist management with scheduling fund pricing accountants, each working 4-day weeks, optimally convering fund pricing assignments over the 5-day investment week. MA34.5 No Title Supplied • Tanu S. Bhatnagar; George Washington University, 403 Monroe Hall, 2115 G St. NW, Washington, DC 20052; tsb@gwu.edu • Srinivas Y. Prasad; George Washington University, 403 Monroe Hall, 2115 G St. NW, Washington, DC 20052; prasad@gwu.edu We address some issues in the design of web-based DSSs. Specifically, an XML-based architecture for distributed decision support is developed. The different types of interactive analyses that such an architecture would enable are illustrated through prototypical examples. # Information Systems III Session: MA35 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Gerardine Desanctis Chair Address: Duke University, Fuqua Sch. of Bus., Box 90120, Durham, NC 27708-0120 Chair E-mail: gd@mail.duke.edu Chair: Chair Address: Chair E-mail: MA35.1 Managing the Emergence of E-Commerce in Russia We assess the prospects for successful e-commerce in Russia's emergent market economy. We explore re-engineering and paradigm shifts in industries, processes, relationships, services and products. Russia lacks mature telecommunications and information infrastructures. However, leaders who overcome some or all of these obstacles enjoy strategic advantages. MA35.2 'Soft' Risks in IS Projects during Planning Stages • Halia M. Senu; University of South Australia, 1 Landscape Crescent Highbury, Adelaide SA, 5089 , Australia; halis_senu@cobweb.com.au • Tricia Vilkinas; University of South Australia, City West Campus, North Terrace Adelaide, Adelaide, SA, 5000 , Australia; Technical or 'hard' aspects can account for a poor 10% of project failure. Over 90% of the cause attributed to IS project failures are due to 'soft' factors. Humans are a risk, and this is often absent from any risk analysis. We propose to define the value of including humans as an identified risk factor when undertaking a risk analysis for the IS project. MA35.3 Sharing Knowledge in Computer-Supported Groups: Some Findings from Real-World Interventions We explore the application of a new format for sharing knowledge in real-world computer-supported groups. It provides support, in the form of facilitator observations and quantitative data analyses, for a claim that this format actually benefits the group members in them being able to share more knowledge. MA35.4 Managing Information Technology-Related Communication: Transactional vs. Relational Approaches • Gerardine Desanctis; Duke University, Fuqua Sch. of Bus., Box 90120, Durham, NC 27708-0120; gd@mail.duke.edu • Izak Benbasat; University of British Columbia, Faculty of Commerce, 2053 Main Mall, Vancouver, BC, V6T 1Z2 , Canada; izak.benbasat@ubc.ca Three types of communication are important to management of IT in future enterprises: communication between technology and people using technology communication between technology designers and technology users and communication between IT units and business units. Using a transactional vs. relational dichotomy, we argue that IT professionals and general managers can and should take active steps to design relationship-based communication... # Operations Management I Session: MA36 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Jack C. Hayya Chair Address: Pennsylvania State University, 303 Beam Bldg., University Park, PA 16802 Chair E-mail: jch@psu.edu Chair: Chair Address: Chair E-mail: MA36.1 Process Management with Uncertain Environmental Limits • Chialin Chen; Wilfrid Laurier University, Sch. of Bus. & Economics, Waterloo, Ontario, N2L 3C5 , Canada; cchen@wlu.ca • George E. Monahan; University of Illinois, Dept. of Business Admin., 1206 South 6th St., #350, Champaign, IL 61820; gmonahan@uiuc.edu Based on a stochastic model of production planning and inventory control, we develop a DSS for process management with both demand and environmental uncertainties. A major finding shows that standards on production wastes may result in transferring environmental risks from one stage of a product's life cycle to another. MA36.2 A Business to Business Bargaining Model with Supply Uncertainty • Haresh B. Gurnani; HKUST, Dept. of ISMT, Clear Water Bay, Kowloon, , Hong Kong; mnharesh@ust.hk • Mengze Shi; HKUST, Dept. of Mktg., Clear Water Bay, Kowloon, , Hong Kong; mkshi@ust.hk A buyer enters into a contract with an unreliable supplier but they have different beliefs on the level of uncertainty in the channel. We first derive the Nash-bargaining solution. Next, we discuss the role of down payments in the contract and analyze incentive-compatibility issues. Finally, we consider a non-symmetric contract that maximizes channel profits under certain conditions. MA36.3 When Does Cross-Training Servers Pay Off? We consider a service system with 2 types of customers and flexible and dedicated servers. Using simulation, we investigate the impact of various system parameters on server mix of flexible and dedicated servers. For a variety of situations, we show that when there is a need for flexibility, a small fraction of flexible servers is sufficient in most situations to minimize the total service and delay costs. MA36.4 Research Issues in Supply Chain Design: Analytic Modeling Approaches • Jack C. Hayya; Pennsylvania State University, 303 Beam Bldg., University Park, PA 16802; jch@psu.edu • Chao-Hsien Chu; Pennsylvania State University, Sch. of Info. Tech., Rider Bldg., University Park, PA 16802; chu@ist.psu.edu • Xin X. He; South Carolina State University, School of Bus., 300 College St. NE, Orangeburg, SC 29117; xhe@scsu.edu We examine the various modeling approaches in supply chain design from traditional mathematical programming to emerging information technologies, such as GAs. These would be compared to simulation and other more practical approaches. Advantages and disadvantages would be outlined and commented upon. # Optimization Techniques V Session: MA37 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Rex K. Kincaid Chair Address: College of William & Mary, Dept. of Mathematics, Williamsburg, VA 23187-8795 Chair E-mail: rrkinc@math.wm.edu Chair: Chair Address: Chair E-mail: MA37.1 withdrawn - author request of 10/29 MA37.2 Performance Measures for Aviation Security Systems Aviation security is an important problem of national interest. We introduce performance measures to assess the effectiveness of airport baggage screening security device systems. These measures are then used to formulate optimization problems for identifying optimal airport baggage screening security device deployments. MA37.3 An Atypical Evolutionary Algorithm for Structural Optimization • Rex K. Kincaid; College of William & Mary, Dept. of Mathematics, Williamsburg, VA 23187-8795; rrkinc@math.wm.edu • Shelley Griffith; College of William & Mary, Dept. of Comp. Sci., Williamsburg, VA 23187-8795; • Jaroslaw Sobieski; NASA, Langley Research Ctr., Hampton, VA 23681; An evolutionary based strategy, BCB, utilizing 2 normal distributions to generate children is presented. A 2-phase approach coupling a local search (either pattern search or a quasi-Newtonsearch) with BCB to improve performance is also examined. Test cases include a hub design problem. MA37.4 Local Search: Are Search Intensity & Diversity Mutually Exclusive? One implicit tenet of local search metaheuristics is that there is a mutually exclusive balance between 2 desirable goals: search exploration, i.e., diversity of areas, and search exploitation, i.e., intensity in a specific area. We suggest that these goals are not mutually exclusive and can indeed be obtained simultaneously. # Industry Applications III Session: MA38 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Bruce R. Feiring Chair Address: Suffolk University, 8 Ashburton Pl., Boston, MA 02108 Chair E-mail: bfeiring@suffolk.edu,, http://saywer.suffolk.edu Chair: Chair Address: Chair E-mail: MA38.1 'The Workbox': Sequencing Final Assembly for Hybrid Pull-ATO Manufacturing at Lucent Technologies We describe a new final assembly sequencing system developed for Lucent Technologies, Shreveport. The system will enable transitioning from a build-to-stock to a near-assemble-to-order paradigm. While many factors affect the feasibility of a pure ATO mode, the workbox is designed to operate in any possible pull/ATO hybrid. MA38.2 Operations Strategy: Convenience Store Location • Hisashi X. Kurata; University of Wisconsin, Sch. of Bus., Milwaukee, WI 53201; • John J. Liu; University of Wisconsin, Sch. of Bus., Milwaukee, WI 53201; jjl@uwm.edu • Jack C. Hayya; Pennsylvania State University, 303 Beam Bldg., University Park, PA 16802; jch@psu.edu We discuss the operations strategy for locating convenience stores. We present a case study of the Unimart convenience stores, State College, Pennsylvania. We apply regression, location analysis and a minimum-distance integer programming to verify the optimality of existing locations. We explain why the results depart from theory in some important aspects. MA38.3 Simulation of a Pull-Push System in an Assembly Line • Mehmet Savsar; Kuwait University, College of Eng. & Petroleum, MIE Dept., PO Box 5969, Safat, 13060 , Kuwait; mehmet@kuc01.kuniv.edu.kw This study deals with simulation modeling and analysis of an electronic assembly line. Weekly demands, which are met from the final inventory of circuit boards, trigger the assembly operations. The objective of simulation is to determine the minimum number of batches in the system, including the WIP and finished circuit board inventory, to assure that a certain percentage of demand is met on time. MA38.4 The Cost of Quality for a Mexican Manufacturer A stochastic modeling procedure is used to obtain the cost of quality for a Mexican subcontractor. This enabled management to analyze costs of internal and external failure costs based upon Juran's cost of quality concept. # AI in Health Care Session: MA39 Date/Time: Monday 08:15-09:45 Type: Invited Sponsor: Track: Cluster: Computational Intelligence Room: Chair: Murali S. Shanker Chair Address: Kent State University, Dept. of MIS, Coll. of Bus., Kent, OH 44242-0001 Chair E-mail: mshanker@kent.edu Chair: Chair Address: Chair E-mail: MA39.1 Machine Learning for Object Detection & Segmentation • W. Nick Street; University of Iowa, MS Dept., 108 Pappajohn Bldg., Iowa City, IA 52242; • Kyoung-Mi Lee; ; Medical decision making relies increasingly on precise and repeatable image analysis. We present a new approach that uses unsupervised learning on deformable shape models to locate and outline objects, improving its performance with continued use. The system then learns to classify the objects based on user feedback. MA39.2 Improving Accuracy of Neural-Network Predictors for Medical Diagnosis • Murali S. Shanker; Kent State University, Dept. of MIS, Coll. of Bus., Kent, OH 44242-0001; mshanker@kent.edu • Michael Y. Hu; ; • Ming S. Hung; ; We present results of using several techniques such as bootstrapping and cross-validation to improving the prediction of neural networks applied to medical diagnosis. Applications to breast-cancer prediction, diabetes diagnosis and discrimination of remote complex tones on pitch perception are presented. # Applied Probability I Session: MA40 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Michael H. Veatch Chair Address: Gordon College, Dept. of Math., Wenham, MA 01984 Chair E-mail: veatch@gordon.edu Chair: Chair Address: Chair E-mail: MA40.1 Warranty Reserve for 2-Dimensional Warranty Policies Warranty reserve analysis is used by producers to decide on the amount of money they need to cover the warranty cost. We consider 3 2-D warranty policies: 2-D renewable free-replacement, 2-D non-renewable free-replacement and 2-D pro rata warranty policies. MA40.2 A Fluid Model for a Processor Sharing Queue • Amber L. Puha; California State University, Dept. of Mathematics, San Marcos, CA 92096; apuha@csusm.edu Formally, a fluid model is a first order, or law of large numbers, approximation of a stochastic system. We propose and analyze a fluid model for the processor sharing queue. It turns out that an appropriate state descriptor for the processor sharing queue is a measure valued process that at each fixed time puts a unit of mass on the residual service time of each job... MA40.3 Nonatomic Total Rewards Markov Decision Processes with Multiple Criteria We consider a Markov decision process with an uncountable state space for which the vector performance functional has the form of expected total rewards. Under the single condition that initial distribution and transition probabilities are nonatomic, we prove that the performance space coincides with that generated by nonrandomized Markov policies. We provide applications of these results to investment and inventory control problems. MA40.4 Explicit Fluid Limits for Multiclass Queueing Networks • Michael H. Veatch; Gordon College, Dept. of Math., Wenham, MA 01984; veatch@gordon.edu Consider the fluid limit of a MQNET with exponential service and inter-arrival times and policies that have a threshold form. Dai showed that these limits exist. We give a method for computing the resulting fluid trajectories and show the connection with the second vector field of Malyshev. # Operational Issues in Global Supply Chain Execution Session: MA41 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: CPMS Track: Cluster: Room: Chair: Susan Rothberg Chair Address: IBM Integrated Supply Chain Group, 3039 Cornwallis Rd., PO Box 12195, RTP, NC 27709 Chair E-mail: seroth@us.ibm.com Chair: Chair Address: Chair E-mail: MA41.1 Complexity & Commonality Issues at the End of a Product Life Cycle for Workstation Manufacturing & Supply Chains • John Konopka; IBM Integrated Supply Chain Group, 3039 Cornwallis Rd., PO Box 12195, RTP, NC 27709-2195; konopka@us.ibm.com We study complexity and commonality issues from a project that looked at the root causes of the end of life manufacturing overage for workstations. We also discuss the results that provide insights into the effects of commonality and complexity on the overage as well as its effect on the product supply chain. MA41.2 Identify the Optimal Machine Attach Rate of Features (Feature Ratio) in BAU Environments & for Special Bids using Statistical Forecasting • Mamnoon Jamil; IBM Integrated Supply Chain, 1000 Atrium Way, Atrium 1, Mt. Laurel, NJ 08054; mamnoon@us.ibm.com In the workstation manufacturing environment, very few statistical studies have been performed to identify the optimal horizons and weights of actual and order backlogs for the automatic feature attach rate (ratio) calculation. We performed a statistical analysis to identify optimal parameters, which produce the most accurate feature ratio in BAU environment... MA41.3 Leveraging Supply Chains over the Internet • Jay Krishnamurthy; IBM Microelectronics Division, 1000 River Rd., Bldg. 964, Essex Junction, VT 05452; jaykrish@us.ibm.com Business to business applications over the Internet provide a powerful mechanism to extend the range of vision across the supply chain partners. At the same time, this also poses newer challenges to the planning and fulfillment engines. We illustrate how companies can use web-based B2B applications to integrate the supply chain and the associated technical and process challenges. MA41.4 Management Issues in Global Supply Chain Execution • Susan Rothberg; IBM Integrated Supply Chain Group, 3039 Cornwallis Rd., PO Box 12195, RTP, NC 27709; seroth@us.ibm.com This discussion will focus on manufacturing methodology and its impact to supply chain execution. We will explore the trade-off between design for logistics and cost of serviceability, with specific emphasis on the management decision making process. # Six Sigma Methods Development & Applications to Manufacturing Processes Session: MA42 Date/Time: Monday 08:15-09:45 Type: Sponsor/Invite Sponsor: Quality, Statistics & Reliability Section Track: Cluster: Reliability & Quality Control Room: Chair: Theodore T. Allen Chair Address: Ohio State University, Dept. of ISE, 1971 Neil Ave., Columbus, OH 43210-1271 Chair E-mail: allen.515@osu.edu,, http://www-iwse.eng.ohio-state.edu/~facultyp/allen.htm Chair: Chair Address: Chair E-mail: MA42.1 Desirability-Based Methods that Address Process Variability & Methods Comparison for Arc Welding Parameter Optimization • Charlie Ribardo; Stewart & Stevenson Inc. Tactical Vehicle Systems, 5000 I-10 West, Sealy, TX 77474; charlie_ribardo@ewi.org We use desirability functions to compare alternative statistical process design methods based on a single relevant arc welding application. The comparison takes into account an expanded set of criteria that include experimental cost, derived model accuracies, complexity of the procedure and final solution quality. The results from this comparison are used to suggest objectives for future methods and comparisons... MA42.2 Roles for Simulation Optimization & Methods Development within the Six Sigma Framework The 6 sigma methodology can be viewed as a broad framework for the deployment of statistical and optimization techniques for manufacturing process improvement. Inside this framework are roles for design of experiments, modeling and optimization. Simulation optimization heuristics that we have developed permit the creation of design of experiments techniques requiring substantially reduced numbers of experimental runs... MA42.3 Supersaturated Designs that Directly Maximize the Probability of Identifying Active Factors • Mikhail Bernshteyn; Ohio State University, Dept. of ISE, 1971 Neil Ave., Columbus, OH 43210-1271; Supersaturated designs and associated analysis methods have been proposed by several authors to identify active factors in situations in which only a very limited number of experimental runs are available. The availability of the proposed designs depends on the existence of certain factorial designs and these designs are not, in general, optimal with respect to the objectives that they were proposed to address... MA42.4 Robust Optimization to Achieve the Appropriate Sigma Level • Waraporn Ittiwattana; The Edison Welding Institute, Columbus, OH 43210; Robust machine design seeks to maximize the performance of a machine, taking into account uncertainty in the 'noise' factors that cannot be controlled. Surprisingly, methodologies that use design of experiments to create approximate models of relevant quality characteristics and simulation optimization to directly minimize the expected loss have received relatively little attention. The advantages of these approaches are obvious... # Linking Technology Transfer & Enterprise Equity Session: MA43 Date/Time: Monday 08:15-09:45 Type: Sponsored Sponsor: Technology Management Section Track: Cluster: Room: Chair: W. Austin Spivey Chair Address: University of Texas, Div. of Mgmt. & Mktg., 6900 North Loop, 1604 West, San Antonio, TX 78249 Chair E-mail: wspivey@utsa.edu Chair: Chair Address: Chair E-mail: MA43.1 A Meta-Review of Tech Transfer across Academic Disciplines from 1989 • W. Austin Spivey; University of Texas, Div. of Mgmt. & Mktg., 6900 North Loop, 1604 West, San Antonio, TX 78249; wspivey@utsa.edu • J. Michael Munson; Santa Clara University, Santa Clara, CA; • William T. Flannery; University of Texas, Div. of Mgmt. & Mktg., 6900 North Loop, 1604 West, San Antonio, TX 78249; flannery@lonestar.utsa.edu Interest in technology transfer across academic disciplines highlights this paper. A computerized search focuses on 1989-2000. The intent is to identify articles cited most often in order to discover propositions that define the core of this area of MOT. Findings imply the need for a multidisciplinary perspective. MA43.2 The Kyoto Protocol: An Analysis of the Costs & Technologies Available for a Municipal Utility • William C. Gunst; University of Texas, MSMOT Program, San Antonio, TX; We research the financial and physical constraints that could impact a utility if federal or state statutes impose additional emission restrictions as a result of the Kyoto and subsequent treaties. MA43.3 Digital Audio Broadcasting • John W. Kosub; University of Texas, MSMOT Program, San Antonio, TX; We provide comprehensive information about the IBOC DAB technology, the barriers to its adoption and its potential rate of adoption. MA43.4 The Semiconductor Industry Enters the Specific Phase • Andrew Black; University of Texas, MSMOT Program, San Antonio, TX; We use Utterback's framework to posit the transition of the semiconductor industry into the specific phase of industrial innovation. # Software Demonstration III Session: MA45 Date/Time: Monday 08:15-09:45 Type: Software Demo Sponsor: Track: Cluster: Room: Chair: Katie Odem Chair Address: Applied Decision Analysis LLC Chair E-mail: katie.odem@us.pwcglobal.com Chair: Chair Address: Chair E-mail: MA45.1 Introducing DPL 5.0: Decision Analysis Software DPL is professional decision analysis software using influence diagrams, decision trees and spreadsheets. Graphical outputs include rainbow diagrams, risk profiles, policy trees and policy summaries. New in DPL 5.0: time-series percentiles, 2-way rainbow diagrams and much more! # Transportation Session: MA46 Date/Time: Monday 08:15-09:45 Type: Contributed Sponsor: Track: Cluster: Room: Chair: Ismail Chabini Chair Address: MIT, Dept. of Civil & Environ. Eng., 77 Massachusetts Ave. Rm. 1263, Cambridge, MA 02139-4307 Chair E-mail: chabini@mit.edu Chair: Chair Address: Chair E-mail: MA46.1 A Continuous Linear Programming Model for Air Cargo Revenue Management • Itir Karaesmen; Carnegie Mellon University, 360 Posner Hall, GSIA, Pittsburgh, PA 15213; itir@andrew.cmu.edu We propose a continuous linear programming model for revenue management in the air cargo industry. Because of the demand and inventories, even a simplified air cargo problem is challenging. We focus on optimization, prove duality for the continuous LP, provide approximations and discuss numerical results. MA46.2 On Cost Allocation in Networks with Threshold-Based Discounting We consider networks in which every 2 nodes can communicate directly but there is a monetary incentive to combine flows from different sources. Namely, the cost of flow through a particular link is discounted if it exceeds a prescribed threshold. We discuss how to allocate the cost of such networks in a manner that stimulates cooperation of users. MA46.3 Algorithms for Route Guidance Generation • Jon Bottom; MIT, Dept. of Civil & Environ. Eng., 77 Massachusetts Ave. Rm. 1263, Cambridge, MA 02139-4307; jbottom@mit.edu • Moshe E. Ben-Akiva; MIT, Dept. of Civil & Environ. Eng., 77 Massachusetts Ave. Rm. 1263, Cambridge, MA 02139-4307; mba@mit.edu • Ismail Chabini; MIT, Dept. of Civil & Environ. Eng., 77 Massachusetts Ave. Rm. 1263, Cambridge, MA 02139-4307; chabini@mit.edu Route guidance is a key component of intelligent transportation systems applications. Guidance generation can be posed in a number of ways, such as fixed point problems, with typically noisy function evaluations. We examine a number of algorithms for solving such fixed point problems, including the classic method of successive averages and a method proposed by Polyak... MA46.4 Shortest Path Algorithms in Continuous-Time Dynamic Networks with Applications to Transportation Networks with Traffic Lights • Ismail Chabini; MIT, Dept. of Civil & Environ. Eng., 77 Massachusetts Ave. Rm. 1263, Cambridge, MA 02139-4307; chabini@mit.edu We present efficient algorithms for variants of shortest path problems in continuous-time dynamic networks. We model the minimum-time shortest path problem in networks with traffic lights as a special case of continuous-time dynamic networks. We show numerical results that demonstrate the efficiency of the developed algorithms for these classes of continuous-time shortest path problems. # Plenary: The Integrated Supply Chain as an Enabler of E-Business Transformation Session: MP44 Date/Time: Monday 10:00-11:00 Type: Plenary Sponsor: Track: Cluster: Room: Chair: Way Kuo Chair Address: Texas A&M University, Dept. of IE, College Station, TX 77843 Chair E-mail: way@tamu.edu Chair: Chair Address: Chair E-mail: MP44.1 Plenary: The Integrated Supply Chain as an Enabler of E-Business Transformation E-business clearly has become 'the next big thing.' As the rate and pace of the e-business phenomenon continue to accelerate - and as projections for business and consumer spending via e-commerce continue to inflate - designing and deploying an end-to-end supply chain model becomes increasingly important for companies large and small. Mr. Jones will discuss the continuing e-business/e-commerce transformation, with a particular emphasis on the supply chain and value chain integration. He will also discuss practical considerations that companies face in developing their e-business strategies. # Integrating Real Options & Decision Analysis Session: MC01 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: Decision Analysis Society Track: Cluster: Room: Chair: Ronald A. Howard Chair Address: Stanford University, Dept. of MS & Eng., Terman Eng. Ctr., Sch. of Eng., Stanford, CA 94305-4023 Chair E-mail: rhoward@stanford.edu Chair: Chair Address: Chair E-mail: MC01.1 Real Options & the Black Scholes Formula: What's Wrong with this Picture? We aim to reinforce the significance and value of real options by weeding out the most pernicious misconception associated with them, which is the presumed connection between real options and the Black Scholes formula, and point towards preferred alternatives for evaluating real options on solid quantitative and intuitive grounds. MC01.2 Development Options • James E. Matheson; Navigant Consulting Inc., 2440 Sand Hill Rd., Menlo Park, CA 94025; jmatheson@sdg.com Most real options situations consist of a period of development of an opportunity leading to a launch event followed by a period of exploitation. Examples range from classic R&D, to creating new businesses, to making movies, to exploring for natural resources. I will present a general paradigm for characterizing development situations and several modeling and solution methods for evaluating them. MC01.3 Real Options in Negotiation & Agreement Design • Mazen A. Skaf; Global Trading, Rapt, Inc., 81 Bluxome St., San Francisco, CA 94107; mazen.skaf@rapt.com I present a framework for uncovering, structuring and evaluating real options in the course of a negotiation and in the life of an eventual agreement. Using a case on technology licensing agreements, I discuss the buyer's reservation line for negotiating contractual real options and the option value of negotiation. # Tutorial: Dynamic Resource Management - Problems, Models & Algorithms Session: MC02 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Room: Chair: Warren B. Powell Chair Address: Princeton University, Dept. of OR & Financial Eng., CASTLE Lab., Princeton, NJ 08544 Chair E-mail: powell@princeton.edu Chair: Chair Address: Chair E-mail: MC02.1 Tutorial: Dynamic Resource Management - Problems, Models & Algorithms • Warren B. Powell; Princeton University, Dept. of OR & Financial Eng., CASTLE Lab., Princeton, NJ 08544; powell@princeton.edu In an era where the rapid interchange of information is growing, the challenge of managing resources efficiently in a dynamic setting is becoming increasingly important. We motivate this area real applications from freight transportation, military operations, vehicle routing, logistics and distribution and management of agricultural commodities. We then provide a modeling vocabulary that merges concepts from optimization, simulation, dynamic programming and control theory... # Retail Inventory Session: MC03 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: MSOM Track: Cluster: Room: Chair: Richard D. Metters Chair Address: SMU, Cox School of Bus., PO Box 750333, Dallas, TX 75275-0333 Chair E-mail: rmetters@mail.cox.smu.edu Chair: Chair Address: Chair E-mail: MC03.1 Managing Inventory with Multiple Products, Lags in Delivery, Resource Constraints & Lost Sales: A Mathematical Programming Approach • John H. Semple; SMU, Cox Sch. of Bus., PO Box 750333, Dallas, TX 72575-0333; jsemple@mail.cox.smu.edu • Brian T. Downs; Aspen Technology, Inc., Chesapeake Supply Chain Div.; • Richard D. Metters; SMU, Cox School of Bus., PO Box 750333, Dallas, TX 75275-0333; rmetters@mail.cox.smu.edu We develop an order-up-to S inventory model that is designed to handle multiple items, resource constraints, lags in delivery and lost sales without sacrificing computational simplicity. We develop nonparametric estimates of these costs and use them in conjunction with linear programming to produce LP policy. MC03.2 Using Information to Improve Retail Product Freshness of Perishables • Michael E. Ketzenberg; University of North Carolina, 1019 Laurel Hill Rd., Chapel Hill, NC 27514; mketz@aol.com We explore the value of information sharing to improve supply chain management of perishables. In a multi-echelon setting, simulation is used to analyze the relative performance of facilities that utilize downstream supply and demand information in replenishment with facilities that only utilize their own demand and inventory information. MC03.3 Breaking Bulk to Improve Retail Space Management • Richard D. Metters; SMU, Cox School of Bus., PO Box 750333, Dallas, TX 75275-0333; rmetters@mail.cox.smu.edu • Michael E. Ketzenberg; University of North Carolina, 1019 Laurel Hill Rd., Chapel Hill, NC 27514; mketz@aol.com • Vicente A. Vargas; Emory University; The benefits of breaking bulk in retail operations are explored. The focus is on the benefits to space management, rather than the reduction in inventory costs. Using data from the grocery industry, results indicate that profitability can be increased substantially - but only if current inventory replenishment practices are changed. # Queueing Applications in Manufacturing Session: MC04 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: MSOM Track: Cluster: Room: Chair: Nico J. Vandaele Chair Address: University of Antwerp, Prinsstraat 13, Antwerp, 2000 , Belgium Chair E-mail: nico.vandaele@ufsia.ac.be Chair: Chair Address: Chair E-mail: MC04.1 Lot-Sizing & Inventory Decisions: An Integrated Approach • Mark L. Spearman; University of Alabama, Coll. of Commerce & Bus. Admin, 366 Alston Hall, Box 870226, Tuscaloosa, AL 35487-0226; mspearman@cba.ua.edu • Silke Kroeckel; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205; silkek@isye.gatech.edu We consider a make to stock system which is replenished with a system having a bottleneck operation that requires significant changeover times whenever we go from one product to another. Our goal is to minimize inventory subject to a service constraint. The problem is very hard as it is not everywhere convex. We provide a practical solution procedure and discuss computational experience. MC04.2 Models for Comparison of Push, Pull & Hybrid Control Strategies for Manufacturing • Ananth Krishnamurthy; University of Wisconsin, Ctr. for Quick Response Mfg., 1513 University Ave., Madison, WI 53706; akrishna@cae.wisc.edu • Rajan Suri; University of Wisconsin, Ctr. for Quick Response Mfg., 1513 University Ave., Madison, WI 53706; suri@engr.wisc.edu Kanban or 'pull' strategies are popular today and considered to be superior to traditional push methods. Using both queueing and simulation models we compare various pull, push, and hybrid control strategies and show that in certain situations the push or hybrid strategies have better performance than kanban/pull. MC04.3 Complex Product Structures in Multi-Stage Capacitated Manufacturing Systems We analyze the performance of multi-product multi-cell manufacturing systems. Each cell is modeled as a queueing network, possibly subject to a workload limit. For each product, a Bill of Material specifies the parts and/or subassemblies to be manufactured or purchased. Commonality of parts may occur, leading to divergent parts routings. We develop integrated stochastic models for these complex systems and use them to investigate the performance of base stock control policies... # Tutorial: Traffic Modeling for Queues & its Impact on Performance Analysis Session: MC05 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: Applied Probability Society Track: Cluster: Room: Chair: Peter W. Glynn Chair Address: Stanford University, Dept. of MS & Eng., Terman Engineering Ctr., Stanford, CA 94305-4026 Chair E-mail: glynn@leland.stanford.edu Chair: Chair Address: Chair E-mail: MC05.1 Tutorial: Traffic Modeling for Queues & its Impact on Performance Analysis • Peter W. Glynn; Stanford University, Dept. of MS & Eng., Terman Engineering Ctr., Stanford, CA 94305-4026; glynn@leland.stanford.edu We discuss 3 different types of traffic environments: short-range dependent with light (marginal) tails, short-range dependent with heavy tails and long-range dependent with light tails. Specifically, we describe the qualitative behavior of a queue fed by such traffic, i.e., heavy-traffic scaling, most likely path to buffer overflow, relaxation time, etc. # Sustainable Transportation Networks Session: MC06 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: Transportation Science Section Track: Cluster: Room: Chair: Anna Nagurney Chair Address: University of Massachusetts, Isenberg Sch. of Mgmt., Dept. of Finance & Op. Mgmt., Amherst, MA 01003 Chair E-mail: nagurney@gbfin.umass.edu Chair: Chair Address: Chair E-mail: MC06.1 A Methodology for the Disaggregate, Multi-Dimensional Measurement of Neighborhood Type It is common in land-use/travel studies to classify neighborhoods as 'traditional' vs. 'suburban.' We develop a methodology for creating continuous scales along which to measure a neighborhood's type - as experienced by an individual resident. The empirical approach offers graphic illustration that the usual binary approach is deficient. MC06.2 The Sequential Travel Forecasting Paradigm was a Counter-Productive Concept • David E. Boyce; University of Illinois, Dept. Civil & Materials Eng., 3073 ERF, 842 West Taylor St., Chicago, IL 60607-7023; dboyce@uic.edu Making intelligent long-term transportation investment choices requires a capability to forecast urban travel. We explore some of the reasons that urban travel forecasting has been relatively unsuccessful during its 50-year history and offer an alternative forecasting paradigm based on OR methods. MC06.3 Dynamics of a Transportation Pollution Permit System with Stability Analysis & Computations • Ding Zhang; SUNY, Sch. of Business, Oswego, NY 13126; zhang@oswego.edu • Anna Nagurney; University of Massachusetts, Isenberg Sch. of Mgmt., Dept. of Finance & Op. Mgmt., Amherst, MA 01003; nagurney@gbfin.umass.edu We develop a dynamic model of a pollution permit system for congested urban transportation networks, which guarantees that the environmental quality standard is achieved. We establish both local and global stability results and propose a discrete-time algorithm for the approximation of the trajectories. Convergence results and numerical examples are provided. MC06.4 Emission Paradoxes & Policies for Sustainability of Transportation Networks • Anna Nagurney; University of Massachusetts, Isenberg Sch. of Mgmt., Dept. of Finance & Op. Mgmt., Amherst, MA 01003; nagurney@gbfin.umass.edu We identify emission paradoxes which can arise in congested urban transportation networks. These paradoxes reveal that the network topology, travel behavior, cost and demand structure must all be taken into consideration in the development of any policies aimed at sustainability of the transportation infrastructure. # INFORMS Case Competition 2000: Finalists 1 & 2 Session: MC07 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: INFORM-ED Track: Cluster: Room: Chair: James J. Cochran Chair Address: Louisiana Tech. University, Dept. of Comp. IS & Anlysis, Coll. of Admin. & Bus., Ruston, LA 71272 Chair E-mail: cochran@cab.latech.edu Chair: Chair Address: Chair E-mail: MC07.1 INFORMS Case Competition 2000: Finalist 1 MC07.2 INFORMS Case Competition 2000: Finalist 2 # Marketing Issues in E-Business Session: MC08 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: E-Business Section Track: Cluster: Room: Chair: Arvind Rangaswamy Chair Address: Pennsylvania State University, Smeal College of Bus. Admin., 707G BAB, University Park, PA 16802-3007 Chair E-mail: arvindr@psu.edu Chair: Chair Address: Chair E-mail: MC08.1 Decomposing the Repeat-Purchase Process into Visiting & Conversion Behavior We focus on the multiple roles that browsing behavior plays in assisting our ability to understand and forecast repeat-purchase patterns for an online merchant. Using data from Media Metrix, we decompose buyer behavior into 2 distinct sub-models: the time between repeat visits to a particular website and the conversion process through which certain visits are associated with purchase transactions... MC08.2 Forecasting within a Virtual Supply Chain at Garden.com Garden.com acknowledges that accurate forecasting is essential to success in the highly competitive e-commerce environment, especially under the conditions of a virtual supply chain. I will describe the flexible forecasting methods that we have developed to provide reliable predictions in the face of high seasonality, diverse product lines and unique seasonal growth patterns. MC08.3 Efficiency, Agency & Customer Satisfaction in Virtual Environments: A Study of Online Investing In virtual marketplaces, the IS function now plays an increasingly important role in managing the interface between the firm and its external customers. While virtual markets are efficient, they also engender problems of agency. We theoretically motivate and empirically validate a model that integrates issues of efficiency and agency in the context of customer satisfaction in virtual settings... MC08.4 Researching Marketing Strategy using the Net • Arvind Rangaswamy; Pennsylvania State University, Smeal College of Bus. Admin., 707G BAB, University Park, PA 16802-3007; arvindr@psu.edu I describe a web-based tool that we have developed to conduct longitudinal strategy studies. The tool enables firms to share sensitive strategy data with each other anonymously and also produces automatic reports. The tool is useful for generating data for academic studies as well as for guiding industry practice. I also describe results from ongoing studies to measure the e-intensity of firms in various industries. # Auction Strategy in Revenue Management Session: MC09 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: Revenue Management Section Track: Cluster: Room: Chair: Timothy J. Flynn Chair Address: University of Alabama, Commerce & Bus. Admin. Coll., Box 870223, Hoover, AL 35487-0023 Chair E-mail: timjflynn99@aol.com,, flynnt@proctr.cba.ua.edu Chair: Chair Address: Chair E-mail: MC09.1 Auction Design for Efficient Must-Run Generation Procurement Given the physical constraints of power flow, must-run generation is a pervasive feature of electricity markets. When its pricing is embedded into the spot market, a substantial degree of market power can be exercised and it is usually evidenced through great variability of intraday bids. We propose a pricing rule that provides incentives for truthful cost revelation... MC09.2 Auctions for Procuring Options • Rakesh V. Vohra; Northwestern University, MEDS Dept., Kellogg Mgmt. Sch., 2001 Sheridan Rd., Evanston, IL 60208-2009; r-vohra@nwu.edu • James Schummer; ; We propose a class of incentive-compatible, efficient auctions for procuring options. Auctions in this class can loosely be described as ex-ante Clark-Groves schemes. We also show that such auctions can be implemented in a computationally efficient way and discuss an application to electricity markets. MC09.3 The English Auction with Differentiated Commodities We propose a simultaneous English auction for multiple goods, when buyers have valuations satisfying the gross substitute (GS) condition. Truthful revelation of demand is a perfect Bayesian equilibrium if the smallest Walrasian prices correspond to Vickrey payments. However, no ascending price auction can reveal sufficient information to implement the Vickrey mechanism if all GS preferences are allowed. # Panel: Research Opportunities in Scheduling Theory & Applications Session: MC10 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Scheduling Room: Chair: Nicholas G. Hall Chair Address: Ohio State University, 301 Hagerty Hall, 1775 South College Rd., Columbus, OH 43210-1144 Chair E-mail: halln@cob.ohio-state.edu Chair: Chair Address: Chair E-mail: MC10.1 Panel: Research Opportunities in Scheduling Theory & Applications • Andreas S. Schulz; MIT, Sloan Sch. of Mgmt. & OR, BLdg. E53-361, 30 Acorn St., Cambridge, MA 02142-1320; schulz@mit.edu • Chelliah Sriskandarajah; University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO4.7, Richardson, TX 75083-0688; • Steef L. van de Velde; Erasmus University, Rotterdam Sch. of Mgmt., PO Box 1738, Rotterdam, 3000 DR , The Netherlands; svelde@fac.fbk.eur.nl Because of its strong theoretical foundations and ability to model a wide variety of important applications, scheduling is among of the most active fields within OR. Emerging research opportunities in scheduling, ranging from theory to applications, will be discussed. Advice about the publication of scheduling work will be provided. # Computational Integer Programming Session: MC11 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Integer Programming Room: Chair: John E. Mitchell Chair Address: RPI, Dept. of Math Sciences, 110 Eigth St., Troy, NY 12180 Chair E-mail: mitchj@rpi.edu Chair: Chair Address: Chair E-mail: MC11.1 On the 'Seymour' Problem: A Case Study of Solving a Hard MIP The 'seymour' problem is one of the most difficult IP instances in MIPLIB. We report on its optimal solution, using a combination of techniques such as preprocessing, decomposition, disjunctive cuts within B&C and parallel processing in the Condor computing environment. MC11.2 Progress on a General Mixed Integer Programming Solver Issues related to the integration and performance of iterative strengthening of cutting planes, primal heuristics and branching schemes into a general MIPSOL will be described. Computational experience with real MIP instances will be reported. MC11.3 Realignment in the NFL • John E. Mitchell; RPI, Dept. of Math Sciences, 110 Eigth St., Troy, NY 12180; mitchj@rpi.edu When the National Football League expands to 32 teams in 2002, it may be realigned into eight 4-team divisions. Minimizing the intradivisional travel distances is a k-partitioning problem, where a graph's vertices are divided into k sets of equal size. We describe a B&C algorithm for the k-partitioning problem.. # Global Supply Chain Management Session: MC12 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Logistics & Supply Chain Management Room: Chair: Herbert Kotzab Chair Address: Copenhagen Business School, Dept. of Op. Mgmt., Solbjerg Plads 3, Frederiksberg, DK-2000 , Denmark Chair E-mail: hk.om@cbs.dk Chair: Chair Address: Chair E-mail: MC12.1 no show • Poul Erik Christiansen; Copenhagen Business School, Dept. of Op. Mgmt., Copenhagen, , Denmark; • Arnold B. Maltz; Arizona State University, PO Box 874706, Tempe, AZ 85287-4076; arnie.maltz@asu.edu MC12.2 Data Integrity: A Critical Measure of Retail Operational Performance • Nicole DeHoratius; Harvard Business School, Morgan Hall T10, Boston, MA 02163; Retail operations rely on getting the right product to the right place at the right time. However, select operational data is often inaccurate causing problems in inventory management, particularly warehouse-to-store fulfillment. We identify the importance of measuring data integrity and suggest a process approach to improving data quality leading to improved service quality. MC12.3 Supply Chain Coordination with Priority Scheduling • Katariina M. Kemppainen; Helsinki School of Economics, Runeberginkatu 14-16, POB 1210, Helsinki, 00101 , Finland; International supply chains often contain several stages that reach for local objectives without considering the impact on system performance. Our study shows how practical priority index-based scheduling mechanisms can be used to coordinate different functions of supply chains. Some applications are discussed. MC12.4 no show • Andreas Otto; Universitat Erlangen-Nurnberg, Lehrstuhl fur Logistik; # Issues in Supply Chain Coordination Session: MC13 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Manufacturing & Logistics Room: Chair: Stephen M. Gilbert Chair Address: University of Texas, CBA 4.202, B6300, Austin, TX 78712 Chair E-mail: steve.gilbert@bus.utexas.edu Chair: Chair Address: Chair E-mail: MC13.1 Incentive Compatible Pricing Mechanisms for Service Differentiated Demand Classes • Vinayak V. Deshpande; Purdue University, 562 Krannert Bldg., West Lafayette, IN 47907; deshpandev@mgmt.purdue.edu • Morris A. Cohen; University of Pennsylvania, The Wharton School, Dept. of OPIM, 1300 SH-DH, Philadelphia, PA 19104-6366; • Karen L. Donohue; University of Minnesota, The Carlson Sch., Dept. of OMS, Minneapolis, MN 55455; We examine pricing mechanisms to coordinate inventory decisions between a supplier and multiple buyers with differentiated service requirements. We develop an incentive compatible pricing scheme which overcomes information asymmetry in the supply chain. We highlight the benefits of this scheme relative to current practice in the service parts industry. MC13.2 Coordination through a Risk-Free Fashionable Item Return Policy • Z. Kevin Weng; University of Wisconsin, Sch. of Bus., Madison, WI 53706; • Scott T. Webster; Syracuse University, Sch. of Mgmt., Syracuse, NY 13244-2130; stwebste@syr.edu A return policy encourages larger order quantities and can increase manufacturer profit. One downside from a manufacturer's perspective is the possibility of very low profit due to high rebate expense. Taking the viewpoint of a manufacturer selling to a single risk-neutral retailer, we develop a return policy that is risk-free. MC13.3 Intermediate Product Design & Replenishment Policies for Managing Variety in the Steel Industry • Diwakar Gupta; University of Minnesota, 125 Mechanical Eng. Bldg., 111 Church St. SE, Minneapolis, MN 55455; • Brian T. Denton; ; • Keith Jawahir; Dofasco Inc., PO Box 2460, Hamilton, Ontario, L8N 3J5 , Canada; keith_jawahir@dofasco.ca Integrated steel mills need to manage inventories of a large number of intermediate products (slabs) with limited warehouse space. We present a 2-stage stochastic programming model for optimizing slab design and setting replenishment policies. The model captures effects of demand, yield and replenishment time uncertainty. MC13.4 Coordination of Pricing & Multiple-Period Production across Multiple Constant Priced Goods We address the problem of jointly determining a production schedule and a vector of prices for a set of products that have seasonal, price-dependent demands. By identifying and exploiting the special structure of the problem, a solution procedure is developed. Through numerical examples, we show the relationship between seasonalities and pricing policies. # Airline Operations & Infrastructure Session: MC14 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: Aviation Applications Section/Transportation Science Section Track: Cluster: Room: Chair: John-Paul Clarke Chair Address: MIT International Center for Air Transportation, Aeronautics/Astronautics Dept., 77 Massachusetts Ave., Cambridge, MA 02139-4307 Chair E-mail: johnpaul@mit.edu Chair: Chair Address: Chair E-mail: MC14.1 An Integer Programming Model to Evaluate Uncertainty Reductions Resulting from Collaborative Decision Making • Michael O. Ball; University of Maryland, Smith Sch. of Business, Inst. for Systems Research, College Park, MD 20742-1815; mball@rhsmith.umd.edu • Thomas Vossen; University of Maryland, Smith Sch. of Business, Inst. for Systems Research, College Park, MD 20742; tvossen@rhsmith.umd.edu The combination of timely cancellation notices and the use of the compression algorithm have reduced the uncertainty of flight arrival times during ground delay programs. We present an analysis, based on an integer programming model, that relates uncertainty reduction to reduction the airborne queue size needed to maintain close to full utilization of an airport's arrival capacity. MC14.2 A Conceptual Design of a Departure Planner Decision Aid • Yiannis Anagnostakis; MIT International Center for Air Transportation, 77 Massachusetts Ave., Cambridge, MA 02139-4307; yianag@mit.edu • John-Paul Clarke; MIT International Center for Air Transportation, Aeronautics/Astronautics Dept., 77 Massachusetts Ave., Cambridge, MA 02139-4307; johnpaul@mit.edu The development of decision support tools for air traffic controllers calls for a thorough understanding of the links and interactions in ATM operations and requires constant evaluation and assessment. Furthermore, the design of a high-level architecture for airport departure management should be based on a thorough analysis of the airport system and understanding of the needs and constraints in current airport operational procedures... MC14.3 Dynamic Ground Delay Program Models • Tasha R. Inniss; Trinity College, Dept. of Math, Washington, DC 20017; innisst@trinitydc.edu • Michael O. Ball; University of Maryland, Smith Sch. of Business, Inst. for Systems Research, College Park, MD 20742-1815; mball@rhsmith.umd.edu Existing stochastic ground holding models implicitly assume that the amount of assigned ground delay is not adjusted over time. We present a new approach that takes into account the dynamic adjustments to assigned delay that are typically made when weather conditions and forecasts change. We demonstrate this methodology using recently derived capacity scenario distributions. MC14.4 An Optimization Model for a Real-Time Flight Scheduling Problem • Goran Stojkovic; GERAD, Ecole Polytech, 3000 Cote-Sainte-Catherine, Montreal, Quebec, H3T 2A7 , Canada; goran@crt.umontreal.ca • Francois Soumis; GERAD, Ecole Polytech., 3000 Cote-Sainte-Catherine, Montreal, Quebec, H3T 2A7 , Canada; soumis@crt.umontreal.ca • Jacques Desrosiers; GERAD, Ecole des HEC, Montreal, Quebec, H3T 2A7 , Canada; jacques@crt.umontreal.ca • Marius M. Solomon; Northeastern University/GERAD, Boston, MA 02115; msolomon@cba.neu.edu When perturbations occur in the day of operation of an airline, the flight schedule may become infeasible and must be updated. We propose an original formulation designed to determine a new flight schedule that fits with the existing aircraft assignment, maintenance schedule, crew schedule and passenger connections. # Tutorial: Digital Business Community - Supply Chains to Market Places Session: MC15 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Room: Chair: Amiya K. Chakravarty Chair Address: Tulane University, AB Freeman Sch. of Bus., New Orleans, LA 70118 Chair E-mail: amiya.chakravarty@tulane.edu Chair: Chair Address: Chair E-mail: MC15.1 Tutorial: The Digital Business Community - Supply Chains to Market Places Companies in a DBC must be integrated seamlessly for quick response to changes in customer preferences, technology and competition. Handoff inefficiencies may creep in among partners from misalignments in information flows, delivery formats and schedules. To ensure knowledge sharing and alignment in business processes, companies may redesign flows of goods, information and funds. How should capabilities, process-alignment, network-flows and system-transparency be determined in such a DBC? ... # Multi-Dimensional Online Auction Design: A Multiple Criteria Perspective Session: MC16 Date/Time: Monday 13:15-14:45 Type: Invite/Sponsor Sponsor: Group Decision & Negotiation Section Track: Cluster: MCDM Room: Chair: Jeffrey E. Teich Chair Address: New Mexico State University, Las Cruces, NM 88001 Chair E-mail: jteich@nmsu.edu Chair: Chair Address: Chair E-mail: MC16.1 withdrawn - chair request of 9/27 • Martin Bichler; Vienna University of Economics & Business Administration, Dept. of Info. Systems, Augasse 2-6, Vienna, 1090 , Austria; martin.bichler@wu-wien.ac.at MC16.2 Combinatorial Auctions with Multi-Dimensional Inexact Bids & Constraints • Gary J. Koehler; University of Florida, Dept. of DIS, Box 117169, Coll. of Bus., 351 Stuzin Hall, Gainesville, FL 32611-7169; koehler@ufl.edu • Joni L. Jones; University of Michigan, 701 Tappan St., Ann Arbor, MI 48109; culpepjj@ufl.edu We present a new way of using an auction mechanism to facilitate B2B negotiations by allowing inexact bidding with multiple evaluative criteria while providing for constraints unique to the environment. We present the proposed model and simplifying heuristics for the combinatorial allocation problem. MC16.3 A Hybrid Multi-Attribute Auction Method Combining Elements of Auctions & Negotiations • Jeffrey E. Teich; New Mexico State University, Las Cruces, NM 88001; jteich@nmsu.edu • Hannele E. Wallenius; Helsinki University of Technology, Dept. of IEM, PO Box 9500, , 02150 , Finland; hwalleni@hkkk.fi • Jyrki Wallenius; Helsinki School of Economics & Business Administration, PO Box 1210, Helsinki, 00101 , Finland; walleni@hkkk.fi • Alexander Zaitsev; Moscow State University, Moscow, , Russia; alex.zaitsev@mtu-net.ru Negotiauction is an Internet-based system developed for either forward or reverse auctions. It can handle complex multi-dimensional auctions, involving multiple attributes, and underlying explicit constraints. Bidders are ranked across multiple bidder attributes, then based on the ranking, 'bid premiums' can be inserted to discriminate among the bidders. MC16.4 Collaborative Agents for Telecommunications Service Level Management • Alex Bordetsky; California State University, TELCOT Institute, 2420 Camino Ramon, Ste. 205, San Ramon, CA 94853; bord@csuhayward.edu Geographically distributed collaboration and adaptation among emerging multimedia satellite-terrestrial services is addressed. An ensemble of cooperating agents are available to help managers to adapt to user and application profiles and consider multiple conflicting QoS criteria. Cooperating agents solve multiple criteria problems of negotiating networking resources through agent committees. # Pricing, Information & Flexibility in Supply Chain Management Session: MC17 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Performance of Supply Networks Room: Chair: Ana Muriel Chair Address: University of Massachusetts, MIE Dept., Amherst, MA 01089 Chair E-mail: ollieatassi@worldnet.att.net,, muriel@ecs.umass.edu Chair: Chair Address: Chair E-mail: MC17.1 Value of Information in Integrated Supply Chain Management • Linda K. Nozick; Cornell University, Sch. of Civil & Environ. Eng., Ithaca, NY 14853-3501; • Mark A. Turnquist; Cornell University, Sch. of Civil & Environ. Eng., Ithaca, NY 14853-3501; • Lynn Truss; General Motors, Enterprise Systems Lab, Global R&D, 30500 Mound Rd., Warren, MI 48090-9055; • Ted Costy; General Motors, Enterprise Systems Lab, Global R&D, 30500 Mound Rd., Warren, MI 48090-9055; • Jeff Tew; General Motors, Enterprise Systems Lab., Global R&D, 30500 Mound Rd., Warren, WI 48090-9055; • June Ma; Cornell University, Sch. of Civil & Environ. Eng., Ithaca, NY 14853; jm255@cornell.edu What information is necessary at what places, and at what times, to optimize the performance of the supply networks. We develop an analytic model that can be used to gain insights into this question. An illustrative example is discussed. MC17.2 An Assemble-to-Order System with Flexible Customers • Seyed Iravani; Northwestern University, Dept. of IEMS, Evanston, IL 60208-3119; • Louis Luangkesorn; Northwestern University, Dept. of IE/MS, Evanston, IL 60208-3119; • David Simchi-Levi; MIT, Dept. of Civil & Environ. Eng., 77 Massachusetts Ave. Rm. 1171, Cambridge, MA 02139; dslevi@mit.edu We consider an assemble-to-order system in which orders consist of essential and non-essential items. A customer (order) is lost when at least one of the essential items of her order is not available. However, some customers are flexible enough to accept a replacement for their essential items (if those items are not available)... MC17.3 Supplier Management in E-Business: Optimal Mix between Long-Term & Spot-Market Procurement • Victor Araman; Stanford University, Dept. of MS & Eng., Stanford, CA 94305; • Jochen Kleinknecht; Stanford University, Dept. of MS & Eng., Stanford, CA 94305; • Edison Tse; Stanford University, Dept. of MS & Eng., Stanford, CA 94305; • Ram Akella; SUNY, Dept. of IE, Excellence in Global Mgmt. Ctr, Buffalo, NY; Internet-based marketplaces represent an additional channel for the procurement of material or components. We study the effects of such a spot market on a manufacturer who previously relied on traditional long-term procurement contracts. We propose very general conditions under which a mixed strategy between those 2 supply channels is optimal... MC17.4 Impact of Production Flexibility on Supply Chain Performance We analyze the impact that introducing flexible manufacturing plants has on supply chain performance in a make-to-order environment. For this purpose, we analytically determine lost sales, inventory levels and transportation costs in both flexible and dedicated plant settings. # Consumer Trust in Internet Stores Session: MC18 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: E-Commerce Room: Chair: Sirkka Jarvenpaa Chair Address: University of Texas, Ctr. Bus., Technology & Law, MSIS Dept, CBA 5.202, Austin, TX 78712-1175 Chair E-mail: sjarvenpaa@mail.utexas.edu Chair: Chair Address: Chair E-mail: MC18.1 Transference as a means of Establishing Trust in World Wide Web Sites • Katherine J. Stewart; University of Texas; Trust transfer is examined as a cognitive process. We explore how hypertext links and evidence that a Web site is associated with a physical store may both increase initial trust in a site through transference. The role of security concerns regarding transactions on the Web is also considered. MC18.2 The Evolution of Consumer Trust in Electronic Commerce: An Integrated Model • V. Srinivasan Rao; University of Texas, San Antonio, TX; • Sirkka Jarvenpaa; University of Texas, Ctr. Bus., Technology & Law, MSIS Dept, CBA 5.202, Austin, TX 78712-1175; sjarvenpaa@mail.utexas.edu Consumer trust in e-commerce is the aggregation of trust in the Internet environment, the individual stores and third-party guarantors. Trust in each component affects trust in other components. An integrated model of the evolution of trust in electronic commerce is developed based on the interdependencies between the 3 components. MC18.3 Engendering Trust in Web Stores: An Experimental Report of the Effect of Four Mechanisms • V. Srinivasan Rao; University of Texas, San Antonio, TX; • Noam Tractinsky; Ben Gurion University, Dept. of IE & Mgmt., Beer Sheva, 84105 , Israel; noamt@bgumail.bgu.ac.il The willingness of consumers to shop on the Internet is dependent on trust in the store. We examine the effectiveness of 4 mechanisms to engender trust in the internet store: the existence of legal assurances, the offer of store guarantees, independent agency reports and reports from others consumers. Preliminary results from a controlled study are discussed. # Dynamic Models of Product Development Session: MC19 Date/Time: Monday 13:15-14:45 Type: Invite/Sponsor Sponsor: Technology Management Section Track: Cluster: Product Development Room: Chair: Edward G. Anderson Chair Address: University of Texas, Dept. of Mgmt., CBA 4.202, Austin, TX 78712 Chair E-mail: edward.anderson@bus.utexas.edu Chair: Chair Address: Chair E-mail: MC19.1 Evolution of Performance under a Deadline Constraint • David N. Ford; Texas A&M University, Civil Eng. Dept., College Station, TX 77843-3136; dford@civilmail.tamu.edu • Nitin Joglekar; Boston University, Sch. of Mgmt., 595 Commonwealth Ave., Boston, MA 02215; joglekar@bu.edu We model the evolution of product performance for a development project under dead line constraint. Performance is accumulated by working on coupled component design tasks followed by system tests. A variety of resource allocation policies are explored to explain the effects of deadlines on dynamics of a classic rework cycle. MC19.2 Designing the Supply Chain: The Impact of Product & Process Modularity Rapidly changing markets make internal investments in knowledge and capacity assets risky, leading firms to depend more on their suppliers. We model internal knowledge and capacity investment decisions as a function of product and process modularity and we examine their impact on outsourcing decisions and overall industry structure. MC19.3 Accelerating Product Development Firms are under increasing pressure to develop new products quickly. This research concurrently addresses 2 of the most common tools for accelerating development processes: overlapping and crashing. We discuss how overlapping and crashing mutually depend on each other, how they impact total development costs, and we derive optimal acceleration policies. # Ensemble Methods for Data Mining Session: MC20 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Computational Intelligence Room: Chair: W. Nick Street Chair Address: University of Iowa, MS Dept., 108 Pappajohn Bldg., Iowa City, IA 52242 Chair E-mail: Chair: Chair Address: Chair E-mail: MC20.1 Ensemble Techniques for Data Mining in Heterogeneous Environments • Joydeep Ghosh; , Dept. of Elect. & Comp. Eng., ENS 516; • K. Tumer; ; In several real-life situations involving large, distributed data sets, it is not possible to centralize data into a single flat file before applying data mining techniques, because of security/privacy concerns, system incompatibilities, etc. We shall present ensemble techniques that are applicable and robust even in such varied and distributed environments. MC20.2 Feature Selection for Ensembles • David Opitz; University of Montana, Comp. Sci. Dept., , MT; In several real-life situations involving large, distributed data sets, it is not possible to centralize data into a single flat file before applying data mining techniques, because of security/privacy concerns, system incompatibilities, etc. We shall present ensemble techniques that are applicable and robust even in such varied and distributed environments. MC20.3 Robust Classification for Imprecise Environments • Foster Provost; NYU, Info. Systems Dept., Stern Sch. of Bus., New York, NY 10027; Target misclassification costs and marginal class distributions rarely can be specified precisely. However, one can build a hybrid classifier that performs as well as the best available classifier (sometimes better) for any target conditions, when optimizing accuracy, expected cost, precision, recall, workforce utilization, etc. MC20.4 Ensemble Techniques in Large-Scale Classification • W. Nick Street; University of Iowa, MS Dept., 108 Pappajohn Bldg., Iowa City, IA 52242; • Changhui Choi; ; This study focuses on ensemble techniques such as voting in large-scale classification tasks. Computational results show that even simple combinations of predictors improve overall accuracy. We explore the selection of training set size, and present a new method that provides some of the advantages of boosting without computational overhead. # Session VI Session: MC21 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: SICUP Room: Chair: Victor J. Milenkovic Chair Address: University of Miami, Dept. of Math & Comp. Sci., PO Box 249085, Coral Gables, FL 33124 Chair E-mail: vjm@cs.miami.edu,, http://www.cs.miami.edu Chair: Chair Address: Chair E-mail: MC21.1 A Tabu Search Approach for the Pallet Loading Problem We present a tabu search-based algorithm for the manufacturer's pallet loading problem. The algorithm starts generating layouts up to 5 blocks and then performs moves that change the number of pieces in the blocks. Such moves can cause deletion and creation of blocks and they are selected based on the number of pieces that can be added to the former layout. MC21.2 withdrawn - chair request of 10/21 • Georg Wichmann; Fraunhofer IML, Joseph-von-Fraunhofer-Str. 2-4, Dortmund, D-44227 , Germany; wichmann@iml.fhg.de • Guenter Dietze; Fraunhofer IML, Joseph-von-Fraunhofer-Str. 2-4, Dortumund, 44227 , Germany; dietze@iml.fhg.de MC21.3 A Tree Search Algorithm for Solving Heterogeneous Single- & Multi-Container Problems In order to solve heterogeneous single- and multi-container loading problems, an algorithm is presented that builds homogeneous blocks where all items have the same orientation. Additional aspects such as load stability and placement restrictions are also taken into account. For benchmarking the test cases of Bischoff & Ratcliff MC21.4 A Lagrangean Relaxation Heuristic for the Manufacturer's Pallet Loading Problem We developed a heuristic method, based on Lagrangean and surrogate relaxation, to solve the manufacturer's pallet loading problem. Such a problem consists in arranging the maximum number of boxes by layer on the pallet, thus optimizing the utilization of the pallet's surface. A reduction method and a Lagrangean heuristic are applied in a subgradient optimization procedure. # Technical & Allocative Efficiency of Health Care Providers using DEA Session: MC22 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: Health Applications Section Track: Cluster: Room: Chair: Yasar A. Ozcan Chair Address: Virginia Commonwealth University, PO Box 980203, 1008 East Clay, Dept. of Health Admin., Richmond, VA 23298-0203 Chair E-mail: ozcan@hsc.vcu.edu Chair: Chair Address: Chair E-mail: MC22.1 no show MC22.2 The Technical Efficiency of Hospitals under a Single Payer System: The Case of Ontario Community Hospitals • Thomas S. Gruca; University of Iowa, Tippie Coll. of Bus., W376 Pappajohn Bus. Bldg., Iowa City, IA 52242-1000; thomas-gruca@uiowa.edu • Deepika Nath; Ernst & Young, 200 Clarendon St., Boston, MA 02116; deepika.nath@ey.com We find no significant differences in efficiency across ownership, size or location. While consistent with Hansmann (1980), our results conflict with prior research. These differences are due to pooling hospitals with and without key outputs in the same sample. Also, hospitals in Ontario are unable to discriminate between payer groups. MC22.3 no show • Jen-Huei Chen; Cheng Gung University, 259 Wen-Hwa 1st Rd., Rao-Yuan, Taiwan, , ROC; m8740003@stmail.cgu.edu.tw • Yu-Chi Tung; National Taiwan University, 19 Su-Chow Rd., Ste. 308, Taipei, Taiwan, , ROC; • I-Chiu Chang; National Chung-Cheng University, Mgmt. School Bldg., Rm. 602, No. 160, San-Hsing Ming-Hsiung, Taiwan, , ROC; misicc@mis.ccu.edu.tw • Shih-Jung Hsiao; National Chung-Cheng University, Mgmt. School Bldg., Rm. 602, No. 160, San-Hsing Ming-Hsiung, Taiwan, , ROC; MC22.4 Technical Efficiency in Renal Dialysis: An Application of Data Envelopment Analysis • Yasar A. Ozcan; Virginia Commonwealth University, PO Box 980203, 1008 East Clay, Dept. of Health Admin., Richmond, VA 23298-0203; ozcan@hsc.vcu.edu • Hacer Ozgen; Virginia Commonwealth University, PO Box 980203, 1008 East Clay, Dept. of Health Admin., Richmond, VA 23298-0203; We examine the issue of technical efficiency in the production of renal dialysis from freestanding providers of dialysis treatments. This is done using an optimization-based method, namely, DEA. We focus on elaborating the sources and levels of inefficiency in dialysis production process. Based on this information, potential cost-savings for the Medicare end-stage renal disease program are illustrated. # Ranking & Selection for Simulation Session: MC23 Date/Time: Monday 13:15-14:45 Type: Invite/Sponsor Sponsor: Simulation Section Track: Cluster: Simulation Room: Chair: W. David Kelton Chair Address: University of Cincinnati, Dept. of QAOM, PO Box 210130, Cincinnati, OH 45221-0130 Chair E-mail: david.kelton@uc.edu Chair: Chair Address: Chair E-mail: MC23.1 Novel Combinations of Screening & Indifference-Zone Procedures • Justin Boesel; The MITRE Corporation, 1820 Dolley Madison Blvd., MS W625, McLean, VA 22102; boesel@mitre.org • Barry L. Nelson; Northwestern University, Dept. of IE/MS, 2145 Sheridan Rd., Evanston, IL 60208; nelsonb@iems.nwu.edu We present novel combinations of statistical screening and 2-stage indifference zone procedures to return the best stochastic system with minimal experimental effort. Essentially, second-stage sampling information taken from superior systems is used to screen out inferior systems more efficiently, reducing the need for additional sampling on inferior systems. MC23.2 Beyond Selection: Why is this System Best? We show how to augment simulation selection experiments with variance attribution analysis, providing insights into why certain systems behave well (or poorly). The results may also highlight modeling assumptions that are critical determinants of system performance, as well as suggest alternative systems worthy of investigation. MC23.3 Comparison of Procedures to Select the Best Bernoulli Population • David Goldsman; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332; sman@isye.gatech.edu We compare a number of procedures to select the Bernoulli population having the largest success probability. The comparison is based upon achieved probability of correct selection and expected number of observations until procedure termination. We also give computer-simulation applications. # Gaining a Competitive Advantage through Distribution Session: MC24 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Warehousing & Distribution Room: Chair: Russell D. Meller Chair Address: Virginia Tech., Dept. of ISE, 250 New Engineering Bldg., Blacksburg, VA 24061 Chair E-mail: rmeller@vt.edu Chair: Chair Address: Chair E-mail: MC24.1 Distribution at Cross Creek Apparel • Russell D. Meller; Virginia Tech., Dept. of ISE, 250 New Engineering Bldg., Blacksburg, VA 24061; rmeller@vt.edu • John Klote; Virginia Tech., Dept. of ISE, 250 New Engineering Bldg., Blacksburg, VA 24061; jkote51@hotmail.com Through an internship at Cross Creek Apparel, interesting research problems were identified. We present one of the problems and illustrate the impact of its solution on Cross Creek. MC24.2 Use of Dynamic Slotting Policies in Forward Pick Areas • Brett Peters; Texas A&M University, 3131 TAMU, Dept. of IE, College Station, TX 77843-3367; bpeters@tamu.edu • Jeffrey S. Smith; Auburn University, Dept. of ISE, 207 Dunstan Hall, Auburn University, AL 36849-5346; jsmith@eng.auburn.edu Increasing product customization and diversification is driving distribution centers to provide smaller, more frequent deliveries. This requirement precipitates changes in picking policies. One such change, dynamic slotting, is described. Based on work with JC Penney, an overview of the strategy and some initial models are presented. MC24.3 Scheduling Methods for a Cross-Docking Distribution Center We will discuss expressions for estimating the loading times of outbound trucks as a function of the inbound truck schedule and the operating parameters, e.g., unloading rate, conveyor speed, etc., of the facility. Illustrations of applying these formulas to scheduling and optimization of the facility will also be given. MC24.4 Incorporating Optimization into a Warehouse Management System • Dan Ockerman; Manhattan Associates, 2300 Windy Ridge Rd., 7th Floor, Atlanta, GA 30339; dockerman@manh.com We present several examples of how we help to optimize the operations in various distribution centers. Example applications include advanced 3-D cubing, vehicle routing, wave optimization and decision execution support systems. # Complementarity & Applications Session: MC25 Date/Time: Monday 13:15-14:45 Type: Sponsored Sponsor: Optimization Section Track: Cluster: Linear Programming & Complementarity Room: Chair: Michael C. Ferris Chair Address: University of Wisconsin, Dept. of Comp. Sci., 1210 West Dayton St., Madison, WI 53706 Chair E-mail: ferris@cs.wisc.edu,, http://www.cs.wisc.edu/~ferris Chair: Chair Address: Chair E-mail: MC25.1 Fracture Propagation using the PATH Algorithm • J.-Cl. De Bremaecker; Rice University, Dept. of Geology/Geophysics, PO Box 1892, Houston, TX 77251-1892; • Michael C. Ferris; University of Wisconsin, Dept. of Comp. Sci., 1210 West Dayton St., Madison, WI 53706; ferris@cs.wisc.edu,, http://www.cs.wisc.edu/~ferris We use the PATH algorithm to compute the stresses and relative displacements on fractures to find their direction of propagation. This is either the direction which maximizes the normal separation on a virtual extension or the one which maximizes the shear stress thereon. MC25.2 Complementarity Approaches to Support Vector Machines The linear support vector machine can be posed as a complementarity problem in a variety of ways. We look at some reformulations and specialized algorithms for solving these problems based on complementarity. Results on a large 60 million variable problem indicate that these approaches are robust and efficient. MC25.3 Economics & Complementarity We look at some new uses of complementarity in the economics literature and show how to formulate and solve these examples using efficient tools from mathematical programming. # Analysis of Queueing Systems Session: MC26 Date/Time: Monday 13:15-14:45 Type: Invited Sponsor: Track: Cluster: Stochastic Models & Applications Room: Chair: Sudha Jain Chair Address: University of Toronto, Dept. of Stats., 100 Saint George St., Toronto, Ontario, M5S 3G3 , Canada Chair E-mail: jainsu@utstat.utoronto.ca Chair: Chair Address: Chair E-mail: MC26.1 Transient Analysis of theM^{X}/M/\infty$Queue • Steve Drekic; University of Waterloo, Dept. of Stats./Actuarial Sci., Waterloo, Ontario, N2L 3G1 , Canada; skrekic@math.uwaterloo.ca • Gordon E. Willmot; University of Waterloo, Dept. of Stats./Actuarial Sci., Waterloo, Ontario, N2L 3G1 , Canada; We demonstrate that the transient distribution of the number in the system in the$M^{X}/M/\infty\$ queue may be computed numerically in a straightforward manner for a wide variety of bulk arrival distributions. In particular, the truncated negative binomial distribution serves as a nice candidate to illustrate the approach.

MC26.2 Waiting Times in State-Dependent Loss Systems
• Jeff Kharoufeh; Pennsylvania State University, Dept. of IME, 310 Leonhard Bldg., University Park, PA 16802;
• Natarajan Gautam; Pennsylvania State University, Dept. of IME, 310 Leonhard Bldg., University Park, PA 16802;

We derive an expression for the probability distribution of holding time for a state-dependent loss system. The service-rate dependence structure is special in that changes in service rate may occur during a service cycle for any server. We apply the results in vehicular traffic flow problems.

MC26.3 Using Cumulant Functions to Obtain System Characteristics for Tandem Queues
• Timothy Matis; Texas A&M University, Zachary Engineering Ctr., College Station, TX 77843-3131;
• Richard M. Feldman; Texas A&M University, Zachary Engineering Ctr., College Station, TX 77843-3131; richf@tamu.edu

We consider a population-constrained tandem queueing model in which arrivals and services are Markovian. Obtaining system characteristics for this model via solutions to Kolmogorov equations is often computationally intractable. A technique for approximating these characteristics using cumulant functions is presented.

MC26.4 A Bayesian Approach for Point Estimation & Testing Procedures in Queueing Systems
• Sudha Jain; University of Toronto, Dept. of Stats., 100 Saint George St., Toronto, Ontario, M5S 3G3 , Canada; jainsu@utstat.utoronto.ca

A conjugate exponential prior has been taken into consideration to develop a Bayesian approach for estimating parameters in a queueing system. The posterior probabilities of point estimates and from these probabilities, the Bayes estimates are derived for specific distributions. A general test procedure for an exponential family of distribution is also developed.

MC26.5 withdrawn - chair request of 10/30
• T. Rao; Institute for Systems Studies & Analyses, Defence R&D Org., Metcalfe House, Delhi, 110 0054 , India; tssr@mailcity.com
• Sudha Jain; University of Toronto, Dept. of Stats., 100 Saint George St., Toronto, Ontario, M5S 3G3 , Canada; jainsu@utstat.utoronto.ca

# Project Management II

Session: MC27
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Willy S. Herroelen
Chair Address: Catholic University of Leuven, Dept. of Applied Econ., Naamsestraat 69, Leuven, 3000 , Belgium
Chair E-mail: willy.herroelen@econ.kuleuven.ac.be
Chair:
Chair E-mail:

MC27.1 Maximizing the Net Present Value of a Project: A Comparison of Different Solution Procedures

Maximizing the net present value of an unconstrained project has become a well-known objective in project scheduling literature. We compare different solution procedures for this problem where both minimal and maximal time lags are considered. These procedures have been coded in C++ and have been validated on benchmark problem sets.

MC27.2 On the Merits & Pitfalls of Critical Chain Scheduling

Critical chain scheduling and buffer management - the direct application of Goldratt's theory of constraints to project management - has recently emerged as a popular approach to project management. We critically review its project scheduling fundamentals and test its scheduling methodology and assumptions through a full factorial computational experiment.

MC27.3 Reactive Project Scheduling: Some Algorithms & a General Framework

Some considerations about the value of sequentially using deterministic schedules in the presence of uncertainty are presented. The reactive project scheduling problem is formally described. Some alternatives for full rescheduling and their efficiency are discussed. Computational experiments have been conducted and will be used to illustrate the results.

# Railroad Empty Equipment Distribution Systems

Session: MC28
Date/Time: Monday 13:15-14:45
Track:
Cluster:
Room:
Chair: Kevin N. Crook
Chair Address: BNSF Railway, 2500 Lou Menk Dr., Ft. Worth, TX 76131
Chair E-mail: kevin.crook@bnsf.com
Chair:
Chair E-mail:

MC28.1 Car Distribution Optimization at Canadian National Railways
• Anshu Pathak; Canadian National Railways, 935 rue de La Gauchetiere W, Montreal, Quebec, H3B 2M9 , Canada; anshu.pathak@cn.ca

The current car distribution process is described, followed by a discussion of how optimization can be beneficial. Improving the accuracy of car supply and demand forecasts is seen as the foundation for successful optimization and efforts in this area are discussed.

MC28.2 Norfolk Southern Empty Equipment Distribution

A description of the current empty railcar distribution system at NS followed by a discussion of where OR efforts are being targeted to improve the process.

MC28.3 Equipment Distribution Optimization at BNSF Railway

We give an overview of the new EDO system recently implemented at BNSF. The discussion will include an overview of the system, a description of the user training process, preliminary results and possible enhancements.

MC28.4 no show
• David W. Bell; CSX Transportation, 500 Water St., Jacksonville, FL 32202; david_bell@csx.com

# Application of Optimization in the Energy Industry

Session: MC29
Date/Time: Monday 13:15-14:45
Track:
Cluster: OR/MS in the Service Industries
Room:
Chair: Shmuel S. Oren
Chair Address: University of California, Dept. of IE&OR, 4135 Etcheverry Hall, Berkeley, CA 94720
Chair E-mail: oren@ieor.berkeley.edu
Chair:
Chair E-mail:

MC29.1 A New Decomposition Methodology Applied to Large-Scale Optimization Problems: Local & Global Behavior
• F. J. Prieto; University Carlos III, Madrid, , Spain;
• F. J. Nogales; University Carlos III, Madrid, , Spain;
• A. J. Conejo; Universidad de Castilla La Mancha, 13071 Ciudad Real, La Mancha, , Spain; aconejo@ind-cr.uclm.es

A decomposition methodology is described and analyzed. The methodology is based on introducing corrections on the iterates generated by the solution process for a related and decomposable problem. Local and global convergence properties of the procedure are discussed. An application to the solution of multi-area optimal power flow problems is described and numerical results are discussed.

MC29.2 Gas Turbine Scheduling with Maintenance Contracts
• Chung-Li Tseng; University of Maryland, Dept. of Civil Engineering, College Park, MD 20742; chungli@eng.umd.edu
• Mariam Zarrabi; University of Maryland, College Park, MD 20742;

We present an optimization problem faced by an owner of gas turbines, who schedules the units to meet long-term power supply contracts and to sell electricity to spot markets for profit. The scheduling is subject to a maintenance contract that restricts total energy (MWh) generated over a year and is based on which the payment of the contract is determined.

MC29.3 Bidding & Scheduling Hydroelectric Generation into Energy & Ancillary Service Markets
• Alva J. Svoboda; Pacific Gas & Electric Co., 77 Beal St., San Francisco, CA 94105; ajsh@pge.com

Hydroelectric plants are well suited to contribute to ancillary service markets because of their capability to offer capacity backed up by limited energy. However, operations of hydroelectric plants participating in both energy and ancillary service markets requires that uncertainties, such as the amount of energy actually dispatched, be modeled in the bidding and scheduling processes.

MC29.4 Numerical Experiments with a Large-Scale Equilibrium Model for the North American Natural Gas System
• Steven A. Gabriel; University of Maryland, Project Mgmt. Program, Dept. of Civil & Environ. Eng., College Park, MD 20742-3021; sgabriel@eng.umd.edu
• Julio Manik; ICF Consulting, 9300 Lee Highway, Fairfax, VA 22031-1207;
• Shree Vikas; ICF Consulting, 9300 Lee Highway, Fairfax, VA 22031-1207;

We present numerical experiments on a large-scale model for the North American natural gas system. This model uses a successive LP approach to solve a nonlinear market equilibrium model. We report the results of several heuristic approaches used to speed up the computations relative to the number of LP subproblems needed.

# Telecommunication & Generalized Network Optimization Algorithms

Session: MC30
Date/Time: Monday 13:15-14:45
Track:
Cluster:
Room:
Chair: Jeffery L. Kennington
Chair Address: SMU, Dept. of Comp. Sci. & Eng., Dallas, TX 75275
Chair E-mail: jlk@seas.smu.edu
Chair:
Chair E-mail:

MC30.1 Optimizing a WDM-Based Optical Network Design
• Eli V. Olinick; SMU, Dept. of Comp. Sci. & Eng., Sch. of Eng. & Applied Sci., Dallas, TX 75275-0122; olinick@seas.smu.edu

We address a difficult design problem for WDM-based optical networks. We propose an optimization model for selecting and sizing a minimum-cost set of optical rings with sufficient capacity to satisfy a given demand pattern among sites on the network. We present an empirical analysis of heuristic solution techniques.

MC30.2 Modular Spare Capacity Planning for Mesh Restorable Networks: Exact & Heuristic Methods

We discuss solution approaches for the problem of allocating spare capacity in a mesh restorable telecommunications network with modularity constraints. Results are presented for a set of realistically-sized test networks.

MC30.3 Integral Multi-Commodity Flow Routing in Telecommunication Networks

The need to find routes corresponding to network demand bundles that should not be split amongst different routes is a typical problem encountered in network routing. The proposed heuristic attempts to find routes in a capacitated, multi-commodity flow environment where each route is a single, integral flow. The approach has applications to restoration algorithms as well as wavelength routing.

MC30.4 Algorithms for Preprocessing Networks: Theory, Implementation & Recommendations

We provide algorithms for modifying a network in order to solve the associated mathematical programming problem faster. Some methods include adding arcs. We compare our implementation of these algorithms to CPLEX Presolve and Aggregator. Recommendations are made concerning when to preprocess a network problem.

# Open Source Software for Operations Research

Session: MC31
Date/Time: Monday 13:15-14:45
Track:
Cluster:
Room:
Chair: Robin Lougee-Heimer
Chair Address: IBM TJ Watson Research Center, PO Box 218, Rte. 134, Yorktown Heights, NY 10598-0218
Chair E-mail: robinlh@us.ibm.com
Chair:
Chair E-mail:

MC31.1 A Primer on Open Source Software
• Donald K. Rosenberg; Stomian Technologies, 919 Monmouth Ave., Durham, NC 22701; donr@stromian.com

Don Rosenbeg, the author of 'Open Source: The Unauthorized White Papers,' will speak on the topic of open source software and its benefits for both researchers and businesses. Don will explain the Open Source Definition, cover how intellectual property and open source co-exist and clarify the difference between open source licenses.

# Telecommunication Systems II

Session: MC32
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Kerem Tomak
Chair Address: University of Texas, CBA 5.202, B6500, Austin, TX 78712
Chair E-mail: k_tomak@yahoo.com,, http://www.bus.utexas.edu/faculty/msis/tomakk
Chair:
Chair E-mail:

MC32.1 Performance of FFT-EMT Telecommunication Systems
• Jose R. Coll; Universidad Central de Venezuela, Apartado 81080, Prados del Este, Caracas, Miranda, 1080 A , Venezuela; jcoll@elecrisc.ing.ucv.ve

Performance of fast Fourier transform discrete multi-tone telecommunication systems, where optimum allocation of bits and spectrum shaping is previously made in order to maximize the data rate, is presented for band-limited channels which can introduce noise, inter-symbol and inter-channel interference. Discrete multi-tone systems achieve higher capacity than single carrier systems.

MC32.2 Channel Occupancy Time of Overflow Calls in Hierarchical Cellular Systems
• Kunmin Yeo; Pohang University of Science & Technology, Dept. of IE, San 31 Hyoja-dong, Kyungbuk, 790-784 , Korea; kunmin@postech.ac.kr
• Chi-Hyuck Jun; Pohang University of Science & Technology, Dept. of IE, San 31 Hyoja-dong, Pohang, 790 784 , Korea; chjun@postech.ac.kr

We focus on the modeling of channel occupancy time of overflow calls in hierarchical cellular systems, which can provide alternate routes for blocked calls. Our model takes into account of the general distribution for cell dwell time. Based on the characterization of call time by an Erlang distribution, the Laplace transform of channel occupancy time distribution is derived...

MC32.3 Path-Based Restoration for Logical Networks: Single Node or Link Failures
• Richard T. Wong; Telcordia Technologies/AT&T Labs., 33 Knightsbridge Rd., PY4-4N316, Piscataway, NJ 08854; rwong1@telcordia.com
• Hanan Luss; Telcordia Technologies, 33 Knightsbridge Rd., PY4-4N227, Piscataway, NJ 08854; hluss@telcordia.com

We first discuss a heuristic for designing survivable networks subject to a single node failure - augmenting capacities along selected cycles. Next, we discuss how this approach can be adapted to designing survivable networks subject to a single link failure (work performed and released at AT&T Labs).

MC32.4 A Genetic Algorithm Approach to a Design of a WDM Optical Network

We apply a GA approach to a virtual topology design of a wide-area WDM optical network. Based on a given physical topology, a virtual topology consisting of a set of optical light paths is constructed. The objective is to minimize the maximal throughput and accommodate on-growing traffic requirements in a timely fashion. Computational results will be presented.

MC32.5 Creating Incentives to Join an Electronic Marketplace

Digital marketplaces are web sites where a group of companies conduct sales and other transactions. There are obvious advantages to it in the form of increased efficiency but are there any associated risks as well? We adress the problem of creating incentives to join an electronic marketplace.

# Production & Scheduling II

Session: MC33
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Mary E. Kurz
Chair Address: University of Arizona, Dept. of SIE, PO Box 210020, Tucson, AZ 85721-0020
Chair E-mail: maryk@sie.arizona.edu
Chair:
Chair E-mail:

MC33.1 A General Heuristic for Production Planning Problems
• Mathieu P. Van Vyve; Universite Catholique de Louvain, CORE, Voie du Roman Pays 34, Louvain-la-Neuve, 1348 , Belgium; vanvyve@core.ucl.ac.be
• Yves Pochet; Universite Catholique de Louvain, CORE, Voie du Roman Pays 34, Louvain-la-Neuve, 1348 , Belgium; ypochet@core.ucl.ac.be

We propose a new general purpose heuristic for solving production planning problems formulated as MIPs and involving capacities, setup costs and times, multiple items, multi-level product structure, etc.This heuristic requires the solution of a sequence of LP and is compared to classic heuristics using standard lot-sizing test problems.

MC33.2 withdrawn - author request of 9/29
• Roberto C. Pontes; Treic Web & OR Systems, Rua Sao Clemente 185, AP 1705 BL 2, Rio de Janeiro, 22260-001 , Brazil; rcvpontes@hotmail.com,, http://www.treic.com.br
• Lucidio Cabral; Universidade Federal do Rio de Janeiro, Caixa Postal 68511, Rio de Janeiro, 22290-140 , Brazil; lucidio@uol.com.br
• Marcone J. Freitas; Universidade Federal do Rio de Janeiro, Caixa Postal 68511, Rio de Janeiro, 22290-140 , Brazil; marcone@hotmail.com

MC33.3 Flexible Flowline Scheduling with Sequence-Dependent Setup Times
• Mary E. Kurz; University of Arizona, Dept. of SIE, PO Box 210020, Tucson, AZ 85721-0020; maryk@sie.arizona.edu
• Ronald G. Askin; University of Arizona, Dept. of SIE, PO Box 210020, Tucson, AZ 85721-0020; ron@sie.arizona.edu

A mixed integer programming model, lower bounds and heuristics for flexible flow lines to minimize makespan are presented. The problem characteristics considered include the number of stages, asymmetric sequence dependent setup times, constant or non-constant number of machines per stage and jobs which may skip stages. Computational results are presented.

# Multicriteria Decision Making I

Session: MC34
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Marina V. Polyashuk
Chair Address: Northeastern Illinois University, 900 West Rand Rd., Apt. A205, Arlington Heights, IL 60004
Chair E-mail: mfogpol@aol.com
Chair:
Chair E-mail:

MC34.1 Feasibility Mapping of Non-Linear Systems

Algorithms have been recently developed for mapping the global feasible space of multi-attribute systems to support decision modeling and selection. We extend the approach to more generalized non-linear systems. Simulations will be presented to show the effectiveness of the method for decision-making problems with unknown preferences between multiple performance attributes.

MC34.2 A Multicriteria Optimization with Unknown Functional Forms of Objectives with an Application to the Allocation of TV Commercials to
• Eiji Takeda; Osaka University, Grad. School of Econ., 1-7 Machikanayama, Toyonaka, Osaka, 560-0043 , Japan; takeda@econ.osaka-u.ac.jp
• Katsuaki Tanaka; Setsunan University, Faculty of Bus. Admin. & Info., Neyagawa, Osaka, 572-8508 , Japan; k-tanaka@kjo.setsunan.ac.jp

It is not uncommon that it is difficult to know the functional forms of objectives in a multicriteria optimization. We present a compromise programming approach to the problem, based on a data envelopment approximation of the empirical data. Using the awareness data of TV commercials, an illustrative application is provided.

MC34.3 Some Results on Binary Preference Relations in Multiple Criteria Optimization
• Marina V. Polyashuk; Northeastern Illinois University, 900 West Rand Rd., Apt. A205, Arlington Heights, IL 60004; mfogpol@aol.com

Some theoretical and methodological aspects of binary preference relations involving several independent criteria will be discussed. Different types of binary relations will be considered, including invariant relations, as well as relations based on relative importance of criteria.

# Information Systems IV

Session: MC35
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Shankar Sundaresan
Chair Address: Pennsylvania State University, 316 Beam Bldg., University Park, PA 16802
Chair E-mail: shankar-s@psu.edu
Chair:
Chair E-mail:

MC35.1 Examining IT Investment Returns using Weill's Theory of Differing Roles: A Micro & Macro Analysis
• Melinda K. Cline; University of North Texas, 24 Skyline Dr., Trophy Club, TX 76262; clinem@unt.edu

We use Weill's theory of differing IT investment roles to investigate returns. We examine more than 5.6 million hours of development on 8,700 projects. Findings confirm and extend the theory by quantitatively and qualitatively demonstrating that IT returns vary in type and magnitude based on managerial objective.

MC35.2 Experimental Results on Assessing Web Site Quality
• William H. Rybolt; Babson College, MS Dept., Babson Park, MA 02457; rybolt@babson.edu
• David P. Kopcso; Babson College, MS Dept., Babson Park, MA 02457; kopcso@babson.edu
• Leo L. Pipino; University of Massachusetts, Coll. of Mgmt., 1 University Ave., Lowell, MA 01854; leo_pipino@uml.edu

The WWW has exacerbated the challenge to organizations of achieving and maintaining high information quality. Using a model of a web-based information system adapted from Mason & Mitroff, we describe an experiment that tests the interactions of the variables in the model and their relationships to quality dimensions.

MC35.3 Why Information Systems Projects are Abandoned: A Leadership & Communication Theory & Explatory Study
• Effy Oz; Pennsylvania State University, 30 East Swedesford Rd., Malvern, PA 19355; effyoz@psu.edu
• John J. Sosik; Pennsylvania State University, 30 East Swedesford Rd., Malvern, PA 19355; jjs20@psu.edu

Data regarding reasons for IS project abandonment were collected from a sample of chief information officers and their immediate subordinates. Factor analysis results provided 5 major factors: lack of corporate leadership, poorly communicated goals/deliveries, inadequate skills and means, poor project management and deviation from timetable/budget. We derived a PLS model showing relationships among these factors...

MC35.4 Impact of Information Infrastructure on Supplier Networks in E-Business
• Shankar Sundaresan; Pennsylvania State University, 316 Beam Bldg., University Park, PA 16802; shankar-s@psu.edu

The information infrastructure for e-business demands effective integrated inter-organizational information systems. Advances in Internet technologies are changing the reach, range and costs of such information systems. We propose game-theoretic economic models to analyze how these changes in information infrastructure may impact the reconfiguration of suppler networks.

# Operations Management II

Session: MC36
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Hosun Rhim
Chair Address: Hanyang University, College of Bus. & Economics, Seoul, 133-791 , Korea
Chair E-mail: hrhim@netial.com
Chair:
Chair E-mail:

MC36.1 no show
• Rose Li; Middlesex University, The Business Sch., The Burroughs, London, NW4 4BT , UK; r.li@mdx.ac.uk
• David Gallear; Middlesex University, The Business Sch., The Burroughs, London, NW4 4BT , UK; d.gallear@mdx.ac.uk

MC36.2 no show
• Gopala Govinda Rajan; Kochi Refineries Ltd., Ambalamugal PO, Ernakulam, Kerala, 682302 , India; ggr@md2.vsnl.net.in

MC36.3 no show
• Xiaohang Yue; University of Texas at Dallas, Sch. of Mgmt., Richardson, TX 75083; xyue@utdallas.edu
• Chelliah Sriskandarajah; University of Texas at Dallas, Sch. of Mgmt., Box 830688, MS JO4.7, Richardson, TX 75083-0688;
• Houmin Yan; Chinese University of Hong Kong, Dept. of SE/EM, Ho Sin Hang Engineering Bldg., Hong Kong, Shatin NT, , PR China; yan@se.cuhk.edu.hk

MC36.4 Competitive Location, Production & Market Selection
• Hosun Rhim; Hanyang University, College of Bus. & Economics, Seoul, 133-791 , Korea; hrhim@netial.com
• Teck-Hua Ho; University of Pennsylvania, The Wharton Sch., Mktg. Dept., 1400 SH-DH, Philadelphia, PA 19104; hoteck@wharton.upenn.edu
• Uday S. Karmarkar; UCLA, Anderson Sch., 110 Westwood Plaza, Box 951481, Los Angeles, CA 90095-1481; uday.karmarkar@anderson.ucla.edu

We investigate how firms should select their production sites, capacities and quantities under rivalry. The formulation addresses multi-market, oligopolistic spatial competition with heterogeneity in production and logistics costs. We analyze the pure-strategy Nash equilibria of the entry game and provide sufficient conditions for the existence of equilibria in the simultaneous entry game...

# Integer Programming I

Session: MC37
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: John K. Karlof
Chair Address: University of North Carolina, Math. & Stats. Dept., Wilmington, NC 28403
Chair E-mail: karlof@uncwil.edu
Chair:
Chair E-mail:

MC37.1 Condition Measures for Integer Programs
• Jorge R. Vera; Catholic University of Chile, Schoo. of Eng., Vicuna Mackenna 4860, Santiago, , Chile; jvera@ing.puc.cl

We present some ideas for defining condition measures in integer programming, as an extension of existing concepts in continuous optimization. Some of these measures are based on the notion of distance to ill-posedness. We show connections with complexity analysis of some combinatorial algorithms.

MC37.2 Solving Multi-Item Capacitated Lot-Sizing Problems with Setup Times by Branch & Cut

We study the MCL problem with setup times. By analyzing submodels obtained from a single time period of MCL, we obtain strong valid inequalities for it. Computational results obtained by applying these inequalities within a B&C algorithm suggest that our contributions represent significant progress in solving the MCL problem.

MC37.3 Local Algorithms in Discrete Programming
• Oleg A. Shcherbina; Crimean Academy of Ecoprotective Building, Dept. of Computer Sci., PB 118, Simferopol, Crimea, 95000 , Ukraine; soa@ecopro.crimea

LAs using a special structure of block matrix of constraints in applied DP problems are effective for solving block DP problems. An LA is a decomposition algorithm. Matrices with block-tree structure are of special interest for us. The possibilities how to expand LA to different special classes of DP problems are considered...

MC37.4 An Integer Programming Solution to Locating Radioactive Waste Sites
• John K. Karlof; University of North Carolina, Math. & Stats. Dept., Wilmington, NC 28403; karlof@uncwil.edu
• Hongtao Xu; University of Alabama, Computer Sci. Dept.;

The 1980 Low-Level Radioactive Waste (LLRW) Policy Act suggests several constraints for choosing LLRW sites. We consider the problem of choosing the sites and optimizing transportation risks. We model this problem as an integer programming problem and develop a B&B algorithm with penalties to solve it.

MC37.5 Role of Two-Period Submodels & Continuous 0-1 Knapsack Relaxations in Solving Multi-Item Capacitated Lot-Sizing Problems with Setup Times
• Yves Pochet; Universite Catholique de Louvain, CORE, Voie du Roman Pays 34, Louvain-la-Neuve, 1348 , Belgium; ypochet@core.ucl.ac.be
• Andrew J. Miller; CORE, 34 Voie du Roman Pays, Louvain-la-Neuve, 1348 , Belgium; miller@core.ucl.ac.be

We introduce valid inequalities for the MCL problem with setup times, a standard production planning model, which are derived by considering a 2-period submodel of MCL and constructing specific 0-1 continuous knapsack relaxations of this submodel. Computational experience suggests that these inequalities significantly help to solve instances of MCL.

# Public Programs/Processes

Session: MC38
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Harold Vazquez
Chair Address: US Air Force Research Laboratory, 1864 4th St., Bldg. 15, Ste. 1, WPAFB, OH 45433
Chair E-mail: harold.vazquez@wpafb.af.mil
Chair:
Chair E-mail:

MC38.1 The Impacts of Privatization on Electric Utilities in the UK
• Toshio Ariu; Central Research Institute of Electric Power Industry, 1-6-1 Ohtemachi, Chiyoda-ku, Tokyo, 100-8126 , Japan; ariu@criepi.denken.or.jp

We have analyzed the management of electric utility companies in the UK before and after privatization from 3-phase corporate evaluation such as management results, resources and activities. Consequently, we should take a lot of factors into consideration when we discuss privatization or deregulation from management aspects.

MC38.2 Pay for Performance in the Federal Government
• Harold Vazquez; US Air Force Research Laboratory, 1864 4th St., Bldg. 15, Ste. 1, WPAFB, OH 45433; harold.vazquez@wpafb.af.mil
• Robert C. Rue; SRA International, 1777 NE Loop 410, Ste. 510, San Antonio, TX 78217; bob_rue@sra.com

The Air Force Research Laboratory recently completed its third year under an experimental civilian personnel management system that links pay increases to performance. We include an overview of the system, a summary of results and highlight of the operations research techniques used to support the experiment.

# New Product Development I

Session: MC39
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Chungsuk Ryu
Chair Address: Georgia Institute of Technology, DuPree Coll. of Mgmt., 755 Ferst Dr., Atlanta, GA 30332-0520
Chair E-mail: gt7091c@prism.gatech.edu
Chair:
Chair E-mail:

MC39.1 Online Customer Communities, Knowledge Co-Creation & New Product Development
• Satish Nambisan; RPI, Lally Sch. of Mgmt., Troy, NY 12180; nambis@rpi.edu
• David Wilemon; Syracuse University, Sch. of Mgmt., Ste. 400, Syracuse, NY 13244; dlwtexas@aol.com

The promise of customers as a resource in product development has not met its potential. One problem is the issue of connectivity, which refers to the communication channels between a firm and its customers. Fortunately, new communication technologies are transforming the relationship between developers and their customers. We provide a framework for examining this potentially powerful tool.

MC39.2 Study of an Effective Design Support Model at the Upper Stage of the Product Design

We propose the design support model to structure conceptual designs with high stability of design quality at the upper stage of product design rationally. We will organize the model by using reformed AHP, in particular, the basic concept of the model is based on axiomatic design theory.

MC39.3 An Integrated Product Planning Model for Developing a Product Family
• Chungsuk Ryu; Georgia Institute of Technology, DuPree Coll. of Mgmt., 755 Ferst Dr., Atlanta, GA 30332-0520; gt7091c@prism.gatech.edu
• Nagesh N. Murthy; Georgia Institute of Technology, DuPree Coll. of Mgmt., 755 Ferst Dr., Atlanta, GA 30332-0520; nagesh.murthy@mgt.gatech.edu

We propose an integrated product planning model that considers marketing, engineering design and manufacturing implications of developing a family of products that offers configurational variety using non-platform components. We identify the set of non-platform components that should be developed to configure a product family that maximizes profits.

# Applied Probability II

Session: MC40
Date/Time: Monday 13:15-14:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Assaf J. Zeevi
Chair Address: Stanford University, Info. Systems Lab, Stanford, CA 94305-9510
Chair E-mail: assaf@isl.stanford.edu,, http://www-isl.stanford.edu/~azeevi
Chair:
Chair E-mail:

MC40.1 On the Structure of Decision Rules in Stochastic Reservoir Optimization

In previous work, we described the structure of optimal policies for hydro-power reservoirs when the natural inflows were assumed unusable until the next planning period. Here, we examine the impact of a more realistic assumption under which the natural inflows are taken into account at a constant rate throughout the current planning period...

MC40.2 A Longitudinal Least Squares Analysis of Mortgage Rates

Mortgages interest rates have gone up and down more sharply over the past 20 years than they had previously since World War II. Interest rates on capital market instruments can be influenced by monthly changes and the longer-term trend changes of economic indicators. The directly affecting factors will be discussed and analyzed in depth for a better understanding of the analysis.

MC40.3 Existence Condition for the Diffusion Approximations of Multiclass Priority Queueing Networks

We extend the work of Chen & Zhang (1999) and establish a new sufficient condition for the existence of the (conventional) diffusion approximation for multiclass queueing networks under priority service disciplines. This sufficient condition relates to the weak stability of the fluid networks and the stability of the high priority classes of the fluid networks that correspond to the queueing networks under consideration.

MC40.4 Dynamic Control & Routing in Large Call Centers: Approximate Analysis via Diffusion Models

We consider a queueing model of a call center providing service to several customer classes. For large call centers, we analyze the problem of dynamically routing customers to agent pools, using diffusion approximations. The asymptotic control problem gives rise to several insights concerning the optimal assignment policy and cross-training of agents.

# Supply Chain Optimization

Session: MC41
Date/Time: Monday 13:15-14:45
Track:
Cluster:
Room:
Chair: Mamnoon Jamil
Chair Address: IBM Integrated Supply Chain, 1000 Atrium Way, Atrium 1, Mt. Laurel, NJ 08054
Chair E-mail: mamnoon@us.ibm.com
Chair:
Chair E-mail:

MC41.1 Mathematical Modeling for Strategic Decision Making: Finally Fulfilling the Promise

A half century ago, many predicted that computer technology would have a profound impact on strategic decision making. The impact, however, was on operations and not on the work of top management. We describe how new tools such as Industria Studio TM are finally addressing the needs of modern strategy formulation.

MC41.2 Ordering Policies under Uncertain Periodic Supply
• Ronald S. Tibben-Lembke; University of Nevada, MGRS/028, Reno, NV 89557; rtl@unr.edu

When a supplier produces on a periodic schedule and the production quantity typically sells out before the next production run, firms must decide carefully how to place their orders. For example, steel plants typically produce on a 6-week rolling schedule. We present theoretical and computational results.

MC41.3 A Simulation Study of Customer Order Fulfillment Policies in a Transportation/Inventory System
• Rajan Batta; SUNY, Dept. of IE, 342 Bell Hall, North Campus, Buffalo, NY 14260-2050; batta@acsu.buffalo.edu
• Mayur Vamanan; SUNY, Dept. of IE, Buffalo, NY 14260-2050;
• Qian Wang; SUNY, Dept. of IE, Buffalo, NY 14260;
• Robert J. Szczerba; Lockheed Martin Systems Integration, Advanced Technology Dept., 1802 St., Rte. 17C, MD 210, Oswego, NY 13827;

We study the effects of different order fulfillment policies on the transportation costs and shortage costs for a manufacturing company with a number of distribution centers to satisfy the demands of geographically dispersed customers. A simulation experiment is used to find the interaction effects between transportation costs, shortage costs and customer order fulfillment policies.

MC41.4 An Application of Latin Hypercube Sampling to Replenishment Simulation
• Daniel D. Wake; NONSTOP Solutions, Inc., 235 Montgomery St., San Francisco, CA 94104; dan_wake@nonstop.com
• Yi Wan; NONSTOP Solutions, Inc.;
• Amit Garg; ;

Replenishment optimization simulations for many tens of thousands of SKUs can be computationally intensive. We have developed a technique utilizing Latin hypercube sampling to allow us to estimate performance for a large population of SKUs based on a small sub-population. We will discuss our business requirements and present results.

# Value Function Approximation in Continuous-State Stochastic Dynamic Programming

Session: MC42
Date/Time: Monday 13:15-14:45
Track:
Cluster: Reliability & Quality Control
Room:
Chair: Victoria C. P. Chen
Chair Address: Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205
Chair E-mail: vchen@isye.gatech.edu
Chair:
Chair E-mail:

MC42.1 Comparison between MARS & ANN Value Function Approximations in Solving an Inventory Forecasting Stochastic Dynamic Program
• Cristiano Cervellera; DIST Universita di Genova, Dept. of Communications & CSS, Via Opera Pia 13, Genova, 16145 , Italy; blackcat@dist.unige.it
• Victoria C. P. Chen; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205; vchen@isye.gatech.edu

We present numerical solutions to inventory forecasting problems using a statistical perspective of the continuous-state SDP optimization model. The accuracy of the OA/MARS SDP solution method (Chen et al. 1999), which employs orthogonal arrays and multivariate adaptive regression splines, is compared to a method employing ANNs to estimate the value function.

MC42.2 Regression Dynamic Programming for Multiple-Reservoir Control
• K-Y. Fan; Cornell University, Sch. of Civil & Environ. Eng., 210 Hollister Hall, Ithaca, NY 14853-3501;
• Christine A. Shoemaker; Cornell University, Sch. of OR/IE, 225 Rhodes Hall, Ithaca, NY 14853-3801; chris@optimus.cee.cornell.edu
• David Ruppert; Cornell University, Sch. of OR/IE, 225 Rhodes Hall, Ithaca, NY 14853-3801; davidr@orie.cornell.edu

In stochastic dynamic programming models of reservoir control problems, the optimal value function is typically computed at grid points of the discretized continuous state space while function values between grid points are approximated. High dimensional problems, e.g. a system of many reservoirs, require prohibitively large computational effort...

MC42.3 Kriging Estimation with a Decision-making Framework for Evaluating Process Technologies in a Wastewater Treatment System
• William J. Welch; University of Waterloo, Dept. of Stats./Actuarial Sci., Waterloo, Ontario, N2L 3G1 , Canada; wjwelch@uwaterloo.ca
• Victoria C. P. Chen; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205; vchen@isye.gatech.edu
• Timothy W. Simpson; Pennsylvania State University, Dept. of Mech. & Nuclear Eng., Dept. IME, 310 Leonhard Bldg., University Park, PA 16802; tws8@psu.edu

Within a wastewater treatment system, there are various technological options that may be chosen at different levels of the process. We consider a system that consists of a liquid process with 11 levels and a solids process with 6 levels. To select among the options, we utilize criteria involving economic cost, size (land area, volume), robustness, odor emissions and global desirability...

# Taking it to the Streets: Health Technology Diffusion

Session: MC43
Date/Time: Monday 13:15-14:45
Track:
Cluster:
Room:
Chair: Ritalinda D'Andrea
Chair Address: , 2193 Grandbury Way, Germantown, TN 38139
Chair E-mail: rdandrea@ix.netcom.com
Chair:
Chair E-mail:

MC43.1 Beautiful Black Health Coalition: Grass Root Technology Diffusion
• Robin J. Womeodu; University of Tennessee, 842 Jefferson, Rm. A607, Memphis, TN 38103;

In order to address the disturbing rise of HIV infection among African-American women, health professionals are borrowing methods of technology diffusion and grass roots community organizing to more effectively communicate HIV prevention methods-this paper discusses a program targeted to African-American beauticians, who are trained to HIV 'health advisors.'

MC43.2 Creating Healthy Communities: Technology Businesses of the Future
• Carol B. Mattick; , 130 East Travis, Ste. 330, San Antonio, TX 78205;

Considering 'high tech-high touch,' we suggest that technology-based businesses, expecting to promote wide acceptance of their products and services, consider societal consequences of these products, especially those related health-promotion and medicine. We add that compliance with product use increases when technology is seen as non-isolating, fostering connection and community.

MC43.3 But then Everyone will Know: HIV Rapid Testing & the Social Consequences of Health Technology Diffusion

In heath issues, technology that conforms with 'best health practices' may not be accepted because of the personal and societal consequences of its use. We explore research techniques and diffusion methods to better identify and intervene in the barriers to acceptance of medical technologies and suggest practices that foster compliance with use.

# Software Demonstration IV

Session: MC45
Date/Time: Monday 13:15-14:45
Type: Software Demo
Track:
Cluster:
Room:
Chair: Mark Wiley
Chair Address: LINDO Systems, Inc., 1415 North Dayton St., Chicago, IL 60622
Chair E-mail: mwiley@lindo.com
Chair: Bjarni Kristjansson
Chair Address: Maximal Software Inc., 2111 Wilson Blvd., Ste. 700, Arlington, VA 22201
Chair E-mail: info@maximal-usa.com

MC45.1 Modeling with LINDO, LINGO & What'sBest!
• Mark Wiley; LINDO Systems, Inc., 1415 North Dayton St., Chicago, IL 60622; mwiley@lindo.com

LINDO Systems will highlight the recent speed enhancements to linear and integer solvers and demonstrate our new Callable Library and line of popular modeling packages: LINDO - our powerful linear and integer programming engine, LINGO - our integrated modeling language with linear/nonlinear solvers and What'sBest! - our large scale linear/ nonlinear spreadsheet solver.

MC45.2 Using the MPL OptiMax 2000 Component Library with VBA for Excel & Visual Basic
• Bjarni Kristjansson; Maximal Software Inc., 2111 Wilson Blvd., Ste. 700, Arlington, VA 22201; info@maximal-usa.com

The OptiMax 2000 is a highly anticipated Component Library (API) designed to take full advantage of the ActiveX Automation component software technologies from Microsoft. Using step-by-step examples, we will demonstrate how to use the OptiMax 2000 to seamlessly integrate MPL into VBA for Excel/Access, Visual Basic, Visual C++, Delphi, Java and the Web, thus easily creating customized end-user applications to solve real-world optimization problems.

# Tutorial: Mining Spatial Data

Session: MC46
Date/Time: Monday 13:15-14:45
Type: Invited
Track:
Cluster:
Room:
Chair: Yupo Chan
Chair Address: University of Arkansas, Systems Eng. Dept., Little Rock, AR
Chair E-mail: novascorp@dayton.net
Chair:
Chair E-mail:

MC46.1 Tutorial: Mining Spatial Data

We show people how to extract useful geographic information in real time. While there are other 'data mining' tutorials, we focus on extracting information to support such applications as facility location, where a facility can be anything from a depot/warehouse to a military target. We also discern trends and patterns on a map, such as spatial signals/communications, weather and pollution plumes...

# Decision Analysis Society Awards Presentation

Session: MD01
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: L. Robin Keller
Chair Address: University of California, Grad. Sch. of Mgmt., 350 GSM, Irvine, CA 92697-3125
Chair E-mail: lrkeller@uci.edu
Chair:
Chair E-mail:

MD01.1 Decision Analysis Society Awards Presentation

Each year, the Decision Analysis Society of INFORMS presents awards for the best publications, the best student paper and the best application of decision analysis. The Society also periodically awards the Ramsey Medal for lifetime contribution; Dr. Detlof von Winterfeldt will be this year's Ramsey award recipient. Current awardees will be honored and will make presentations related to their work.

# Tutorial: Theoretical Foundations of Systems Engineering

Session: MD02
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster:
Room:
Chair: George A. Hazelrigg
Chair Address: National Science Foundation, 4201 Wilson Blvd., Arlington, VA 22230
Chair E-mail: ghazelri@nsf.gov
Chair:
Chair E-mail:

MD02.1 Tutorial: Theoretical Foundations of System Engineering
• George A. Hazelrigg; National Science Foundation, 4201 Wilson Blvd., Arlington, VA 22230; ghazelri@nsf.gov

Systems engineering involves decision making with respect to all aspects of the system. Good systems engineering demands adequate generation of system alternatives and proper selection of the best alternatives from among this set. Selection is the domain of decision theory. We will discuss the underlying concepts of decision making under uncertainty applied to engineering systems...

# Research Related to Telephone Call Centers

Session: MD03
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: Robert Shumsky
Chair Address: University of Rochester, Simon Sch., Rochester, NY 14627
Chair E-mail: shumsky@simon.rochester.edu
Chair:
Chair E-mail:

MD03.1 Staffing with Employee Learning & Turnover
• Noah F. Gans; University of Pennsylvania, OPIM Dept., The Wharton School, Philadelphia, PA 19104-6366; gans@wharton.upenn.edu
• Yongpin Zhou; University of Washington, Dept. of MS, Box 353200, Seattle, WA 98195-3200; yongpin@wharton.upenn.edu

What is the value of tracking detailed learning and turnover information when setting staffing levels? We develop and analyze a model to answer the question. Numerical results suggest that how quickly people learn and how much flexible capacity is available are important determinants of the value of more detailed information.

MD03.2 Admission Control of a Shared Service Facility
• Lerzan E. Ormeci; EURANDOM, PO Box 513, Eindhoven, 5600 MB , The Netherlands; ormeci@eurandom.tue.nl
• Apostolos N. Burnetas; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106-7235; atb4@po.cwru.edu

Many call centers serve different classes of customers distinguished by different profitability, volume of calls, and service expectation. We develop and analyze a queuing model that serves 2 classes of customers using dedicated and shared facilities to develop insight for cross-training of call center operators and dynamic assignment of calls.

MD03.3 Optimal Allocation of Tasks to Call Centers
• Reynold E. Byers; University of California at Irvine, Grad. Sch. of Mgmt., 144 Sea Country Ln., Las Flores, CA 92688; byers@uci.edu
• Gregory Dobson; University of Rochester, Simon Sch., Rochester, NY 14627; dobson@simon.rochester.edu
• Robert Shumsky; University of Rochester, Simon Sch., Rochester, NY 14627; shumsky@simon.rochester.edu

We develop models for allocating tasks to telephone call centers, taking into account queueing effects, wages, training, and the fixed costs associated with each center. The models also consider the benefits of blending a variety of job types, such as inbound service, outbound sales, and responding to email.

# Supply Chain Applications in E-Commerce

Session: MD04
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: S. David Wu
Chair Address: Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015-1582
Chair E-mail: david.wu@lehigh.edu
Chair:
Chair E-mail:

MD04.1 Integrating Pricing, Production & Distribution Strategies within the Supply Chain

The increasing popularity of e-commerce is driving businesses to develop new supply chain management strategies that capitalize on the advantages of e-business. We consider the integration of decisions such as pricing, production and distribution within the supply chain, made possible by an e-commerce environment.

MD04.2 The Value of Information in the Supply Chain: An Application in the Telecommunications Industry

Expanded information availability is only the first step in integrating supply chains - core processes need to be re-engineered to ensure performance improvement. We discuss a case study in the telecommunication equipment manufacturing industry, with special attention to the model used to assess operational improvement along the supply chain.

MD04.3 The Stochastic Inventory Routing Problem

Recently, we presented a computational approach for the stochastic inventory routing problem with direct deliveries. We present an approach for the case with more general routing. Computational results are presented, including results on instances obtained from industry.

MD04.4 Transaction Paradigms in the Electronic Market-Driven Supply Chain
• S. David Wu; Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015-1582; david.wu@lehigh.edu

We consider coordination mechanisms in the e-supply chain through the modeling of electronic market transactions. The focus of our explorations is on the following paradigms: auction and electronic intermediaries, negotiation and proactive equilibrium and market coalitions and alliance mechanisms. We will provide an overview of these paradigms.

# Dynamic Control for Stochastic Networks I

Session: MD05
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: Sunil P. Kumar
Chair E-mail: kumar_sunil@gsb.stanford.edu
Chair: Ruth J. Williams
Chair Address: University of California, Dept. of Math., 9500 Gilman Dr., La Jolla, CA 92093-0112
Chair E-mail: williams@russel.ucsd.edu

MD05.1 A Multi-Class Queue in Heavy Traffic with Throughput Time Constraints: Asymptotically Optimal Dynamic Controls

Consider a queueing system with multiple classes of jobs, each having its own renewal input process, service time distribution, revenue contribution and maximum allowed throughput time. A system manager must decide whether or not to accept new jobs as they arrive and also the order in which to serve jobs that are accepted...

MD05.2 Dynamic Scheduling of a Parallel Server System with Complete Resource Pooling
• Ruth J. Williams; University of California, Dept. of Math., 9500 Gilman Dr., La Jolla, CA 92093-0112; williams@russel.ucsd.edu
• Steven L. Bell; University of California, Dept. of Math., 9500 Gilman Dr., La Jolla, CA 92093-0112; slbell@math.ucsd.edu

We consider a parallel server queueing system with flexible scheduling capabilities. Under a complete resource pooling condition, a continuous review threshold control policy is proposed and it is shown to be asymptotically optimal in the heavy traffic limit.

MD05.3 Dynamic Control for Stochastic Networks: A Comparison of Fluid vs. Brownian Approximations
• Constantinos Maglaras; Columbia Business School, 409 Uris Hall, 3022 Broadway, New York, NY 10027-6902; c.maglaras@columbia.edu

A promising approach for designing control policies for stochastic networks is based on analysis of fluid or Brownian approximating models and translation of the corresponding controls through, for example, tracking or discrete-review policies. We contrast the solutions extracted from these 2 model approximations through simple examples.

MD05.4 A New Numerical Method for Solving Brownian Control Problems
• Sunil P. Kumar; Stanford University, Grad. School of Business, 518 Memorial Way, Stanford, CA 94305-5015; kumar_sunil@gsb.stanford.edu
• Muthukumar Muthuraman; Stanford University, Sci. Computing & Comp. Math., Dept. of Comp. Sci., Stanford, CA 94305-5015; mkumar@stanford.edu

We present a new method for numerically solving Brownian control problems. We adapt nonlinear finite element methods to numerically solve the Hamilton-Jacobi-Bellman equation associated with the Brownian control problem. The solution to this partial differential equation is then used to construct an optimal control for the Brownian system.

Session: MD06
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: Pitu B. Mirchandani
Chair Address: University of Arizona, The ATLAS Ctr., SIE Dept., Tucson, AZ 85721
Chair E-mail: pitu@sie.arizona.edu
Chair:
Chair E-mail:

MD06.1 SMART-RHODES: A Self-Learning Traffic Adaptive Controller
• Pitu B. Mirchandani; University of Arizona, The ATLAS Ctr., SIE Dept., Tucson, AZ 85721; pitu@sie.arizona.edu
• David E. Lucas; University of Arizona, The ATLAS Ctr., SIE Dept., Tucson, AZ 85721; lucas@sie.arizona.edu
• Steve Nobe; University of Arizona, The ATLAS Ctr., SIE Dept., Tucson, AZ 85721; san@sie.arizona.edu
• Wenji Wu; University of Arizona, The ATLAS Ctr., SIE Dept., Tucson, AZ 85721; wenji@sie.arizona.edu

SMART-RHODES, a self-learning real-time traffic adaptive signal control system, provides dynamic, optimal signal phasing that considers the natural stochastic variations in demand. By incorporating self-tuning algorithms to update demand-dependent parameters, such as turning proportions and queue discharge rates, the system continuously provides adaptive phasing, removing the need for manual tuning/calibration.

MD06.2 Computational Aspects of On-Line Traffic Control using Beowulf Clusters
• Pengcheng Zhang; Purdue University, Sch. of Civil Eng., West Lafayette, IN 47907-1284;
• Srinivas Peeta; Purdue University, Sch. of Civil Eng., 1284 Civil Eng. Bldg., West Lafayette, IN 47907-1284;

The computational aspects of implementing various models associated with on-line traffic control in the high-performance Beowulf environment are investigated. Benchmark tests involving shortest path and simulation components are analyzed. Some related issues such as load balancing and data communication are addressed.

MD06.3 Real-Time Signal Priority for Transit Vehicles
• Pitu B. Mirchandani; University of Arizona, The ATLAS Ctr., SIE Dept., Tucson, AZ 85721; pitu@sie.arizona.edu
• Anna Knyazyan; University of Arizona, SIE Dept., Tucson, AZ 85721;
• Wenji Wu; University of Arizona, The ATLAS Ctr., SIE Dept., Tucson, AZ 85721; wenji@sie.arizona.edu

Phase durations are given by real-time optimization (RHODES) that considers all vehicle-delays, passenger-counts in transit vehicles and transit-schedule status (if the vehicle is late or early). In the phase optimization, the vehicle 'weight' increases with the number of passengers and vehicle lateness. Simulations indicate that passenger delays are decreased.

MD06.4 A 0-1 Mixed Integer Linear Programming Formulation for the Traffic Signal Control Problem
• Wei Hua Lin; Virginia Polytechnic Institute & State University, Dept. of Civil & Environ. Eng., Blacksburg, VA 24061; whlin@vt.edu

The proposed 0-1 MILP formulation for the traffic signal optimization problem captures delay, stops and physical queues. It handles fixed and variable cycles, constrained by minimum and maximum green duration. The number of integer variables is equal to the number of intersections multiplied by the optimization duration.

# INFORMS Case Competition 2000: Finalists 3 & 4

Session: MD07
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: James J. Cochran
Chair Address: Louisiana Tech. University, Dept. of Comp. IS & Anlysis, Coll. of Admin. & Bus., Ruston, LA 71272
Chair E-mail: cochran@cab.latech.edu
Chair:
Chair E-mail:

MD07.1 INFORMS Case Competition 2000: Finalist 3

MD07.2 INFORMS Case Competition 2000: Finalist 4

# Online Auctions

Session: MD08
Date/Time: Monday 15:00-16:30
Track:
Cluster: Bidding
Room:
Chair: Paulo B. Goes
Chair Address: University of Connecticut, Operations & Info. Mgmt., Sch. of Bus., Storrs, CT 06269
Chair E-mail:
Chair:
Chair E-mail:

MD08.1 Understanding Revenue Drivers in E-Auctions
• Kumar Mehta; University of Illinois, Dept. of IDS, Chicago, IL 60607; kmehta1@uic.edu

Utilizing the advantages borne out of the new medium of communication and transaction, new business models and sales channels have accompanied the explosive growth of Internet economy. Departing from earlier approach of opening a retail front on the web that essentially inherited similar business and revenue models, businesses are increasingly adopting dynamic pricing mechanisms as a means for revenue maximization...

MD08.2 An Analysis of Business to Consumer On-Line Auctions
• Ravi Bapna; Northeast University, Coll. of Bus. Admin., Boston, MA 02115; r.bapna@nunet.neu.edu
• Paulo B. Goes; University of Connecticut, Operations & Info. Mgmt., Sch. of Bus., Storrs, CT 06269;
• Alok Gupta; University of Connecticut, Op. & Info. Mgmt., Storrs, CT 06269; alok@sba.uconn.edu

We overview several research approaches that we have taken to analyze the increasingly popular multi-unit progressive auctions in the B2C marketplace: analytic modeling, empirical analysis of real web data, laboratory experiments and simulation. The focus has been to provide a broad understanding of the factors that influence the design of the mechanism from revenue-maximization and allocative efficiency perspectives.

MD08.3 Incompletely Specified Combinatorial Auction: Heuristic Solution Methodology to a Combinatorial Optimization
• Gary J. Koehler; University of Florida, Dept. of DIS, Box 117169, Coll. of Bus., 351 Stuzin Hall, Gainesville, FL 32611-7169; koehler@ufl.edu
• Joni L. Jones; University of Michigan, 701 Tappan St., Ann Arbor, MI 48109; culpepjj@ufl.edu

The incompletely specified combinatorial auction is a mechanism that allows for the submission of bids that guide, rather than explicitly specify, the choice of goods to fulfil buyer's needs. To facilitate achieving a satisfying solution to the problem in real time, we present simplifying heuristics for the combinatorial allocation problem.

MD08.4 A Dynamic Pricing Mechanism for Determining Optimal Capacity & Service Mix in Quality-of-Service Environments
• Ravi Bapna; Northeast University, Coll. of Bus. Admin., Boston, MA 02115; r.bapna@nunet.neu.edu
• Paulo B. Goes; University of Connecticut, Operations & Info. Mgmt., Sch. of Bus., Storrs, CT 06269;
• Alok Gupta; University of Connecticut, Op. & Info. Mgmt., Storrs, CT 06269; alok@sba.uconn.edu

Converging digital technologies of audio and video streaming require a certain level of service quality for acceptable performance. Internet content providers offering webcasting of special events are examples of such firms in B@C markets. Companies that deploy commercial video-on-demand servers in high-bandwidth digital cable markets are examples of such firms in B@B markets...

# Tutorial: Revenue Management & Dynamic Pricing

Session: MD09
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster:
Room:
Chair: E. Andrew Boyd
Chair Address: PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006
Chair E-mail: aboyd@prosrm.com
Chair:
Chair E-mail:

MD09.1 Tutorial: Revenue Management & Dynamic Pricing
• E. Andrew Boyd; PROS Revenue Management, 3223 Smith St., Ste. 100, Houston, TX 77006; aboyd@prosrm.com

RM has been employed with great success in the airline, hotel and car rental industries for many years. We introduce the basic concepts of RM and discuss some of the many new industries where RM is being employed. We also discuss the relationship with dynamic pricing, auctions and exchanges.

# Complex Scheduling Problems

Session: MD10
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Scheduling
Room:
Chair: Peter J.S. Brucker
Chair Address: University of Osnabrueck, Fac. of Math. & Info., Osnabrueck, D-49069 , Germany
Chair E-mail: peter.brucker@mathematik.uni-osnabrueck.de
Chair:
Chair E-mail:

MD10.1 Solving a Chemical Batch Scheduling Problem by Local Search

A chemical batch scheduling problem is modeled in two different ways as a complex scheduling problem. Both models are used to solve the batch scheduling problem in a 2-phase tabu search procedure. The method is tested on real world data.

MD10.2 Batching Identical Jobs
• Philippe Baptiste; University of Technology of Compiegne, HeuDiaSyC, Ctr. Recherches de Royallieu, Compiegne Cedex, 60205 , France; philippe.baptiste@hds.utc.fr

We study the problems of scheduling jobs, with different release dates and equal processing times, on serial batching machines and on parallel batching machines. We show that in both environments, for a large class of objective functions, the problems are polynomially solvable by dynamic programming.

MD10.3 A Complex Scheduling Problem in Telecommunication
• Christian Prins; University of Technology of Troyes, Dept. GSI Laboratory LOSI, 12 rue Marie Curie, Troyes Cedex, 10010 , France; prins@univ-troyes.fr
• Christelle Gueret; University of Technology of Troyes, Dept. GSI Laboratory LOSI, 12 rue Marie Curie, Troyes Cedex, 10010 , France;

Algorithms are proposed to schedule tasks with a common deadline and renewable resources on a set of processors. A processor can execute tasks requiring a resource only if it is equipped with that resource. The goal is to install a minimum number of resource units, while matching the common deadline.

MD10.4 Preemption Can Make Parallel Machine Scheduling Problems Hard
• Peter J.S. Brucker; University of Osnabrueck, Fac. of Math. & Info., Osnabrueck, D-49069 , Germany; peter.brucker@mathematik.uni-osnabrueck.de
• Svetlana Kravchenko; University of Osnabrueck, Fac. of Math. & Info., Osnabrueck, D-49069 , Germany;

It is shown that the problem of scheduling jobs with identical processing times on identical parallel machines with the objective of minimizing the weighted number of late jobs is NP-hard if preemption is allowed. However, on the other side the corresponding nonpreemptive problem can be solved polynomially.

# Branch & Cut Algorithms

Session: MD11
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Integer Programming
Room:
Chair: Roger Z. Rios
Chair Address: Universidad Autonoma de Nuevo Leon, Systems Eng. Program, AP 111-F, Cd. Universitaria, San Nicolas, NL, 66450 , Mexico
Chair E-mail: roger@uanl.mx,, http://osos.fime.uanl.mx/~roger/
Chair:
Chair E-mail:

MD11.1 The Allocation & Scheduling of Trainers & Trainees to Sites & Sessions
• Stephanie R. Earnshaw; Research Triangle Institute, 3040 Cornwallis Rd., PO Box 12194, RTP, NC 27709-2194; earnshaw@rti.org
• Salah E. Elmaghraby; North Carolina State University, Program in OR, Box 7913, Raleigh, NC 27695-7913; elmaghra@eos.ncsu.edu
• Yahya Fathi; North Carolina State University, Dept. of IE, PO Box 7906, Raleigh, NC 27695-7906; fathi@eis.ncsu.edu

We treat the problem of the optimal allocation/location/scheduling of trainers and trainees in face-to-face interview surveys. We propose and ILP model, which is efficiently solved via a B&C/heuristic algorithm. The computational results demonstrate the applicability of the

MD11.2 Matrix Reduction Techniques for Dense MIPs & their Application to Radionuclide Implants for Prostate Cancer

Matrix reduction techniques are developed to tackle dense MIP instances. The procedure involvesdecomposing the constraint matrix, solving the reduced system, and restoring the solutions for the original instance. Computational results related to radionuclide implants for prostate cancer will be presented.

MD11.3 A Branch & Cut Algorithm for the Vehicle Routing Problem
• Ted Ralphs; Rice University, Dept. of Comp. & Applied Math, 6630 Rodrigo St., Houston, TX 77007-2045; ted@mailzone.com
• Leonid Kopman; Caliper Corporation; kopman@caliper.com
• Leslie E. Trotter; Cornell University, Dept. of OR/IE, Ithaca, NY 14853-3801; ltrotter@orie.cornell.edu
• Bill Pulleyblank; IBM TJ Watson Research Center, PO Box 218, Yorktown Heights, NY 10598; wrp@watson.ibm.com

We will discuss recent results with a parallel B&C solver for the VRP. Various aspects of the algorithm will be discussed with a focus on separation. Specifically, we will discuss efforts to 'borrow' separation technology from efforts to solve the TSP using a paradigm that has application

# Designing the Supply Chain

Session: MD12
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Logistics & Supply Chain Management
Room:
Chair: Vedat Verter
Chair Address: McGill University, Faculty of Mgmt., 1001 Sherbrooke St. West, Montreal, Quebec, H3A 1G5 , Canada
Chair E-mail: verter@management.mcgill.ca
Chair:
Chair E-mail:

MD12.1 The Optimization of Manufacturing Network Structures
• Guy Desaulniers; Ecole Polytechnic de Montreal, GERAD, 3000 ch. Cote-Ste-Catherine, Montreal, Quebec, H3T 2A7 , Canada;
• Alain Martel; Universite Laval, CENTOR, Fac. des Sci. & Admin., Quebec, Quebec, G1K 7P4 , Canada; alain.martel@fsa.ulaval.ca
• Marc Paquet; Universite Laval, CENTOR, Fac. des Sci. & Admin., Quebec, Quebec, G1K 7P4 , Canada;

With current globalization trends, several companies must redesign their manufacturing networks, that is determine the number, location and mission of their plants as well as the technology to use in these plants. We propose a mathematical programming model which can be used to support these decisions. We also show how the model can be solved using Benders decomposition.

MD12.2 Modeling & Design of Flexible & Robust Supply Chains
• Marc Goetschalckx; Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205;
• Tjendera Santoso; Georgia Institute of Technology, Sch. of ISyE, 765 Ferst Dr., Atlanta, GA 30332-0205;

The design of strategic supply chains is many times based on uncertain data. We present a design framework, models, and solution algorithms to design flexible and robust supply chains that explicitly incorporate this uncertainty. Some preliminary computational experience will also be reported.

MD12.3 Contract-Based Tactical Planning of a Supply Chain
• Sophie D'Amours; Universite Laval, CENTOR, Quebec, Quebec, G1K 7P4 , Canada; damour@gmc.ulaval.ca
• Benoit Montreuil; Universite Laval, CENTOR, Quebec, Quebec, G1K 7P4 , Canada;

We address the decision process of dynamically planning a product-based supply chain. Our model minimizes operating costs and plans material flows in the supply chain over a rolling horizon. The result of the model is used to define long, mid and short terms contracts with suppliers, subcontractors, business units and

MD12.4 Evaluating Manufacturing Strategies in Production-Distribution Systems: A Continuous Approximation
• Abdullah Dasci; McGill University, Fac. of Mgmt., 1001 Sherbrooke St. West, Montreal, Quebec, H3A 1G5 , Canada; adasci@management.mcgill.ca
• Vedat Verter; McGill University, Faculty of Mgmt., 1001 Sherbrooke St. West, Montreal, Quebec, H3A 1G5 , Canada; verter@management.mcgill.ca

We will present a model for simultaneous optimization of market areas in a 2-echelon production- distribution system where demand as well as the other parameters is represented as continuous functions over the entire market. We will present the closed form solution and use of the model to evaluate

# Game Theoretic Models for Manufacturing Logistics

Session: MD13
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Manufacturing & Logistics
Room:
Chair: Suleyman Karabuk
Chair Address: Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015
Chair E-mail:
Chair:
Chair E-mail:

MD13.1 A Nash Game for Due-Date Coordination between Manufacturing & Marketing
• Murat Erkoc; Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015;
• S. David Wu; Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015-1582; david.wu@lehigh.edu

We analyze the global and competitive optimization of due-date quotation and capacity management in a make-to-order environment. Motivated by the fact that the competition degrades overall performance of the firm, we design a Nash game that induces incentives for the two parties to coordinate and work towards the system optimum.

MD13.2 Coordinating Delivery Schedules between a Buyer & Supplier in the Supply Chain
• Kadir Ertogral; Lehigh University, Dept. of IME, 200 West Packer Ave., Bethlehem, PA 18015;
• S. David Wu; Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015-1582; david.wu@lehigh.edu

We propose a 2-phase coordination mechanism for the delivery scheduling between a buyer-supplier pair in the supply chain. Phase I identifies a delivery schedule that is system optimal while accommodating the interests of both parties. Phase II constitutes a profit sharing scheme between the parties based on a bargaining model.

• Hakan Golbasi; Lehigh University, Dept. of IME, 200 West Packer Ave., Bethlehem, PA 18015;
• S. David Wu; Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015-1582; david.wu@lehigh.edu

We analyze supplier-buyer coordination in a supply chain. Expanding traditional EOQ type models, we develop lead-time based decision models representing the cost structures of the supplier and the buyer. We analyze equilibrium solution conditions taking into consideration queuing effects and WIP inventory costs.

• Mingzhou Jin; Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015-1582;
• S. David Wu; Lehigh University, Dept. of IMSE, 200 West Packer Ave., Bethlehem, PA 18015-1582; david.wu@lehigh.edu

Buyer-Centric eCommerce auction has been widely adopted to lower procurement costs. While suppliers benefit from increased sales opportunities, their margins often suffer. We propose a game-theoretic mechanism that allows suppliers to form coalitions to improve cost efficiency. We show that this mechanism improves overall market efficiency.

# Advances in Stochastic Integer Programming

Session: MD14
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Integer Programming
Room:
Chair: Nikolaos V. Sahinidis
Chair Address: University of Illinois, Dept. of Chem. Eng., 600 South Mathews Ave., Urbana, IL 61801-2906
Chair E-mail: nikos@uiuc.edu,, http://archimedes.scs.uiuc.edu
Chair:
Chair E-mail:

MD14.1 A Branch & Price Algorithm for Stochastic IP & its Application to Simple Integer Recourse Problems

Stochastic programming problems are known to be amenable to decomposition methods. We will discuss the application of a B&P method for stochastic IP problems and specializations that are possible when the problem has a structure called simple integer recourse.

MD14.2 An Asymptotically Optimal Heuristic for Capacity Expansion under Uncertainty

We address a class of multi-period, multi-facility expansion problems under uncertainty that we formulate as multi-stage stochastic integer programs. We devise a decomposition-based heuristic that constructs integral solutions from the LP relaxation solution. The heuristic is proven to be asymptotically optimal in the number of planning periods.

MD14.3 Solving Stochastic Integer Programs with IBM's OSL Stochastic Extensions
• Gyana Parija; IBM TJ Watson Research Center, Math Sci. Dept., PO Box 218, Yorktown Heights, NY 10598; parija@us.ibm.com
• Alan J. King; IBM TJ Watson Research Center, Math Sci. Dept., PO Box 218, Yorktown Heights, NY 10598; kingaj@us.ibm.com

The C-API of IBM's OSL Stochastic Extensions presents a set of new functions to handle integer variables present in stochastic LPs. We describe how stochastic integer programs can be effectively solved by combining well known MIP preprocessing techniques such as branching rules, prioritization of sets, etc. with a wide variety of decomposition strategies available in the stochastic programming literature...

# Scalable Statistical Techniques for Quality Assurance

Session: MD15
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Scalable Statistical Techniques for Quality Assurance
Room:
Chair: Nong Ye
Chair Address: Arizona State University, Dept. of IE, Box 875906, Tempe, AZ 85287
Chair E-mail: nongye@asu.edu
Chair:
Chair E-mail:

MD15.1 Scalable Clustering Techniques for Quality Assurance of Information Systems
• Nong Ye; Arizona State University, Dept. of IE, Box 875906, Tempe, AZ 85287; nongye@asu.edu
• Xiangyang Li; Arizona State University, Dept. of IE, Box 875906, Tempe, AZ 85287;

Intrusions into information systems compromise the security, i.e. availability, integrity and confidentiality, etc., of information systems. Intrusions must be detected at an early stage to prevent damages to information systems. Decision trees techniques have been used to learn intrusion signatures and to classify incoming events in information systems as normal or intrusive...

MD15.2 A Robust Clustering Method based on L1-Median

A robust clustering method, the K-medians method, based on the L1-median is introduced. Examples show that the new clustering method can outperform both the traditional K-means method and the SAS's K-means method (PROC FASTCLUS) in the sense that the new method can appropriately identify clusters in a given data set, whereas the other 2 may fail to do so.

MD15.3 Camouflaging Computer Networks based on Traffic Statisical Information

We discuss how to apply statistical traffic modeling techniques in network security. In particular, we study the development of countermeasures for traffic analysis and denial of service for both wired networks and wireless ad-hoc networks. For traffic analysis attacks, we examine traffic stuffing algorithms that can effectively mask the actual operational modes of mission critical applications without compromising guaranteed quality of service...

MD15.4 On-Line Monitoring of SMD Assembly with Variable Sampling Intervals
• J. Rene Villalobos; Arizona State University, Dept. of IE, Box 875906, Tempe, AZ 85287-5906; rene.villalobos@asu.edu
• Luis F. Munoz; Arizona State University, Dept. of IE, Box 875906, Tempe, AZ 85287-5906;

The introduction of real-time automated inspection systems onto the factory floor has accentuated the need for monitoring tools capable of performing real-time analysis of the large amounts of data generated by these systems. For example, AVI systems are capable of determining a component's presence, position and placement angle in real-time...

# PROMETHEE MCDA Outranking Method

Session: MD16
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: MCDM
Room:
Chair: J. P. Brans
Chair Address: University of Brussels, Dept. of Stats & OR, Poeinlaan 2, Brussels, B-1050 , Belgium
Chair E-mail: jpbrans@vub.ac.be
Chair:
Chair E-mail:

MD16.1 The PROMETHEE Methodology
• J. P. Brans; University of Brussels, Dept. of Stats & OR, Poeinlaan 2, Brussels, B-1050 , Belgium; jpbrans@vub.ac.be

We present an introduction to MCDA and describe requisites for appropriate methods. PROMETHEE basic data are provided. Additional information includes generalised criteria, outranking graph, outranking flows, PROMETHEE I and II, sensitivity tools, walking weights, profiles of actions, applications and software demonstrations.

MD16.2 The PROMETHEE Associated Visual Interactive Module
• J. P. Brans; University of Brussels, Dept. of Stats & OR, Poeinlaan 2, Brussels, B-1050 , Belgium; jpbrans@vub.ac.be

Decomposition of the PROMETHEE II net flow. The GAIA plane. Visualisation of the actions. Visualisation of the criteria. The decision stick. The PROMETHEE decision axis. PROMETHEE VI: The human brain. Sensitivity tools. PROMETHEE V: MCDA with additional constraints. Applications and software demonstrations.

MD16.3 Mathematical Properties of PROMETHEE
• Philippe G. H. Vincke; Universite Libre de Bruxelles, CP 210/01, Boulevard du Triomphe, Brussels, 1050 , Belgium; pvincke@smg.ulb.ac.be

PROMETHEE is an outranking method devoted to the determination of a ranking in a set of alternatives which are evaluated on several criteria. We analyse this method in terms of mathematical properties and underlying assumptions.

# Supply Chain Operational Strategies

Session: MD17
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Performance of Supply Networks
Room:
Chair: Anthony D. Ross
Chair E-mail:
Chair:
Chair E-mail:

MD17.1 Product Line Selection & Pricing under the Multinomial Logit Choice Model
• Jennifer K. Ryan; Purdue University, Sch. of IE;
• Goker Aydin; Stanford University, Sch. of MSE, Stanford, CA 94305;

We consider a retailer's product line selection and pricing problem under cannibalization and the MNL choice model. We describe the structure of the optimal product line and the optimal prices. We obtain several useful insights into the relationship between the optimal set of products, their 'quality' or attribute levels and their optimal prices.

MD17.2 Coordinated Scheduling & Inventory Abatement
• F. Barry Lawrence; Texas A&M University, ETID Dept., College Station, TX 77843-3367; lawrence@entc.tamu.edu
• Powell Robinson; Texas A&M University, Grad. Sch. of Bus., Dept. of IOM, College Station, TX 77843-4217; probinson@cgsb.tamu.edu

In batch manufacturing systems, a family of items may incur both major and minor setup costs during production. There is a major setup cost when one or more members of the product family are produced and a minor setup cost for each individual item produced. In such situations, the coordinated manufacture of the individual items is economically attractive...

MD17.3 Productivity Analysis for Supply Networks: Measuring Temporal Performance
• Anthony D. Ross; Michigan State University, Business Sch., East Lansing, MI;

We present an approach for identifying best practice decision making units in a multi-product supply network operation. The availability of complete operational data often presents many challenges when used by researchers and management analysts. The data is often incomplete, ill-conditioned or not fully dimensioned. A translation invariant DEA model is formulated...

MD17.4 Supply Chain Partnering & Resource Sharing in the Petroleum Industry
• Powell Robinson; Texas A&M University, Grad. Sch. of Bus., Dept. of IOM, College Station, TX 77843-4217; probinson@cgsb.tamu.edu
• Anthony D. Ross; Michigan State University, Business Sch., East Lansing, MI;

Based on the importance of SCM and partnering to the logistics function and the growing debate and discussion of the concept in this industry, we present a timely demonstration and discussion of the resource sharing concept. As a backdrop, we are using the results from a logistics study. The objectives are to describe the upstream supply chain...

# Industry Standards in E-Commerce Supply Chains

Session: MD18
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: E-Commerce
Room:
Chair: Ramnath Chellappa
Chair Address: University of Southern California, Bridge Hall 401D, Los Angeles, CA 90089-0809
Chair E-mail: chellapp@rcf.usc.edu
Chair:
Chair E-mail:

MD18.1 Rethinking Supply Chains for E-Business

We draw on a field effort with an IT industry consortium to uncover two aspects of the emerging supply chain e-business model: process integration and collaborative knowledge sharing. The 'orchestrated e-Process' model is characterized by groups of business partners electronically orchestrating around the customer based on collaborative process partnering.

MD18.2 The Specification of Process Standards for Electronic Business Interfaces

There are several efforts underway to allow enterprises to inter-operate and exploit fast changing market opportunities through market exchanges. We provide a framework for understanding the scope of these efforts. Using an industry effort example, we illustrate the issues and challenges involved in setting up process specifications.

MD18.3 Disincentives for Standards Formation in E-Markets
• Ramnath Chellappa; University of Southern California, Bridge Hall 401D, Los Angeles, CA 90089-0809; chellapp@rcf.usc.edu

While standards in general create greater social welfare in markets, firms often create certain impediments to sustain competition. In particular, the existence of third party integrators and centralized decision making create disincentives for standards formation. We argue that the enterprise resource planning software markets have suffered from this phenomena and explore the migration of ERP firms towards standards.

# Product Development Process Management

Session: MD19
Date/Time: Monday 15:00-16:30
Track:
Cluster: Product Development
Room:
Chair: Steven D. Eppinger
Chair Address: MIT, Sloan School of Mgmt., Rm. E53-347, Cambridge, MA 02139
Chair E-mail: eppinger@mit.edu
Chair:
Chair E-mail:

MD19.1 Ownership Structures in the Auto Industry: A Property Rights Perspective

We examine the interaction between physical asset ownership and access to critical resources as it affects the investment incentives of a manufacturer and supplier in development of automotive parts. We find part design complexity to be a key determinant of the wide range of contracting choices observed empirically.

MD19.2 The Effects of Product Architecture on Technical Communication in Product Development
• Manuel E. Sosa; MIT, Sloan Sch. of Mgmt., Cambridge, MA 02139; msosa@mit.edu
• Steven D. Eppinger; MIT, Sloan School of Mgmt., Rm. E53-347, Cambridge, MA 02139; eppinger@mit.edu
• Craig M. Rowles; Pratt & Whitney Aircraft, East Hartford, CT 06108; rowlescm@pweh.com

Management of product architecture knowledge by the development organization provides important competitive advantage for established firms facing architectural innovation. We study how the combination of product architecture and organizational structure determines technical communication in development teams. We illustrate our approach by analyzing the development process.

MD19.3 Product Information Evolution & Sensitivity in Product Development: Bi-Directional Changes & Causes of Evolution

Krishnan's discrete 2-phase product development project model identifies evolution and sensitivity as important descriptors of product information. We use a continuous model to relax an important assumption about overlapping initial work and iteration, adds the return of changes to the upstream phase, and explores the causes of evolution.

MD19.4 Multiple Project Management using the DSM Method

We consider multiple development projects that share a common set of resources. Within this setting, the availability of predecessor information is not sufficient to ensure that tasks will be executed on time. We address the availability of resources through a project portfolio plan that constrains task execution based on both information and resource availability.

# Optimization in Data Mining

Session: MD20
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Computational Intelligence
Room:
Chair: Erin J. Bredensteiner
Chair Address: University of Evansville, Dept. of Math, Evansville, IN 47722
Chair E-mail:
Chair:
Chair E-mail:

MD20.1 Feature Selection in Clustering
• W. Nick Street; University of Iowa, MS Dept., 108 Pappajohn Bldg., Iowa City, IA 52242;
• YongSeog Kim; ;
• Filippo Menczer; University of Iowa, MS Dept., S320 PBB, Iowa City, IA 52242;

We formulate the feature subset selection problem in unsupervised learning as a multicriteria optimization problem and search the combinatorial solution space using a specialized genetic algorithm. This approach lets the user explore trade-offs among different Pareto-optimal solutions, and also detects high-quality solutions for both feature subsets and number of clusters.

MD20.2 Separating Surfaces & Decision Trees
• Ahmad Moghrabi; Carleton University, Systems & Computer Eng., 1125 Colonel by Dr., Ottawa, Ontario, K1S 5B6 , Canada;
• John W. Chinneck; Carleton University, Systems & Computer Eng., 1125 Colonel By Dr., Ottawa, Ontario, K1S 5B6 , Canada; chinneck@sce.carleton.ca

Classification accuracy depends on both the algorithm for finding the separating surfaces and on the method used to construct the decision tree. We present methods in which these processes interact: the method of placing the separating surface differs depending on the current status of the developing decision tree.

MD20.3 Unlabeled Data Classification by Support Vector Machines
• Glenn Fung; University of Wisconsin, Computer Sci. Dept., 1210 West Dayton St., Madison, WI 53706;
• Olvi L. Mangasarian; University of Wisconsin, Computer Sci. Dept., 1210 West Dayton St., Madison, WI 53706-1685;

An SVM classifies partially labeled datasets by assigning unlabeled data to one of two classes so as to maximize the separation between the two classes. For totally unlabeled data an oracle labels a small percentage of the data chosen by k-median clustering and then the partially labeled procedure is used.

MD20.4 Duality & Geometry in Support Vector Machine Classifiers
• Erin J. Bredensteiner; University of Evansville, Dept. of Math, Evansville, IN 47722;
• Kristin P. Bennett; RPI, Math Sciences Dept., 110 8th St., Troy, NY 12180;

In SVM classification, maximizing the margin between two linearly separable sets is equivalent to finding the two closest points in the convex hulls. We extend this geometric argument to the inseparable case using reduced convex hulls. The effect of the choice of parameters on the solution becomes geometrically clear.

# Session VII

Session: MD21
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: SICUP
Room:
Chair: Christine L. Valenzuela
Chair Address: Cardiff University, Dept. of Comp. Sci., PO Box 916, Cardiff, CF2 3XF , UK
Chair E-mail: christine@cs.cf.ac.uk
Chair:
Chair E-mail:

MD21.1 A Global Optimization Method for the 3-Dimensional Packing Problem
• Loris Faina; University of Perugia, Dipt. Matematica e Informatica, Via L. Vanvitelli 1, Perugia, 06123 , Italy; faina@unipg.it

We introduce a new geometrical model which reduces the 3-dimensional packing problem to a finite enumeration scheme; a very efficient algorithm is derived. Several tests prove the validity of the algorithm, in particular, a numerical estimate of the asymptotic performance bound is given.

MD21.2 Towards a Systematic Treatment of Constraints of 3-Dimensional Packing Problems
• Hermann Gehring; University of Hagen, Dept. of Bus. Informatics, Hagen, 58084 , Germany;
• Andreas Bortfeldt; University of Hagen, Dept. of Bus. Informatics, Hagen, 58084 , Germany; andreas.bortfeldt@fernuni-hagen.de

We discuss the modeling and algorithmic treatment of often occuring constraints of 3-D packing problems. A data format for 3-D packing problems including different constraints is proposed and experiences gained with the calculation of constrained 3-D packing problems are reported.

MD21.3 Solving the Linear Programming Relaxation of Cutting & Packing Problems: A Hybrid Simplex Method/Subgradient Optimization Procedure

We present an improved method for solving the LP relaxation of the cutting stock problem. The method is based on the relationship between column generation and Lagrange relaxation. We test our procedure on generated data sets and compare it with the traditional column generation approach.

MD21.4 A New Genetic Algorithm for Solving Optimally the Two-Dimensional Cutting Problem
• Loris Faina; University of Perugia, Dipt. Matematica e Informatica, Via L. Vanvitelli 1, Perugia, 06123 , Italy; faina@unipg.it
• F. Mori; University of Perugia, , , Italy;

We introduce a new GA, based on the method of the zones, for finding an optimal solution of a 2-dimensional rectangular packing problem. The algorithm has been tested with many examples from the literature and from random data. Results are very encouraging in comparison with some known GAs. Finally, our algorithm is able to reconstruct perfect rectangular puzzles with more than 20 pieces.

# Recent Issues in Medicare Payment Systems

Session: MD22
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: Peri H. Iz
Chair Address: Health Care Finance Administration, Strategic Plng. & Evaluation, 7500 Security Blvd, MS C31926, Baltimore, MD 21244
Chair E-mail: piz@hcfa.gov,, izperi@iamdigex.net
Chair:
Chair E-mail:

MD22.1 Simulating Outlier Policy Parameters for the Medicare Home Health Prospective Payment System
• Philip G. Cotterill; Health Care Finance Administration, HCFA/OSP, 7500 Security Blvd., MS C32128, Baltimore, MD 21244; pcotterill@hcfa.gov

The Medicare Home Health Prospective Payment System pays for 60-day episodes of care. Additional payments for unusually expensive cases (outlier payments) are subject to specific criteria and aggregate constraints. We discuss the methods used to estimate the parameters that satisfy the policy criteria and constraints.

MD22.2 Risk Adjusting Capitation Payments to Health Plans that Disproportionately Enroll Frail Medicare Beneficiaries
• Gerald F. Riley; Health Care Finance Administration, Office of Strategic Plng., 7500 Security Blvd, MS C32017, Baltimore, MD 21244; griley@hcfa.gov

Medicare policymakers are concerned about the adequacy of diagnosis-based risk adjusters for establishing payments for health plans that disproportionately enroll frail beneficiaries. Two risk adjustment models were evaluated with respect to their ability to predict Medicare costs for groups defined by institutional status and functional impairments.

MD22.3 Developing a Geographic Payment Adjuster for Medicare Managed Care Payments
• Edgar A. Peden; Health Care Finance Administration, OSP/REG/DPR, 7500 Security Blvd, MS C31926, Baltimore, MD 21244; epeden@hcfa.gov

Based on the current service by service spending patterns of managed care providers, current health care financing administration geographic payment indexes for Medicare services are combined with other federal geographic indexes to build a county level geographic payment index for managed care plans. The result is then compared to the simpler geographic index currently being used to adjust Medicare managed care payments.

# Operational Modeling in Semiconductor Wafer Fabrication

Session: MD23
Date/Time: Monday 15:00-16:30
Track:
Cluster: Simulation
Room:
Chair: Neal Pierce
Chair Address: Motorola, Advanced Products R&D Lab., 3501 Ed Bluestein Blvd. MD K10, Austin, TX 78721
Chair E-mail: neal.pierce@motorola.com
Chair:
Chair E-mail:

MD23.1 withdrawn - chair request of 10/23
• Nipa Patel; Advanced Micro Devices, 5204 E. Ben White Blvd., MS 612, Austin, TX 78741; nipa.patel@amd.com

MD23.2 Dealing with the Challenges of Modeling a Constant WIP Fab
• James Berry; Motorola, Advanced Products R&D Lab., 3501 Ed Bluestein Blvd. MD K10, Austin, TX 78721; james.berry@motorola.com
• Neal Pierce; Motorola, Advanced Products R&D Lab., 3501 Ed Bluestein Blvd. MD K10, Austin, TX 78721; neal.pierce@motorola.com
• Rick McKiddie; Motorola, Advanced Products R&D Lab., 3501 Ed Bluestein Blvd. MD K10, Austin, TX 78721; rick.mckiddie@motorola.com

A constant WIP fab is more difficult to simulate in comparison to a pre-determined start rate fab. For a pre-determined start rate fab, the start rate is already established and included in the model as input. With a constant WIP factory, WIP allocations are followed and a customer group must be under their allocation in order to start a lot.

MD23.3 Value-Based Dispatching for Semiconductor Wafer Fabrication
• Neal Pierce; Motorola, Advanced Products R&D Lab., 3501 Ed Bluestein Blvd. MD K10, Austin, TX 78721; neal.pierce@motorola.com
• Tanju Yurtsever; Motorola, Advanced Products R&D Lab., 3501 Ed Bluestein Blvd. MD K10, Austin, TX 78721; tanju.yurtsever@motorola.com

The semiconductor industry has primarily emphasized lot dispatching decisions to achieve cycle time goals and customer on-time delivery satisfaction. We present a prototype lot dispatching system with the objective of VBD that incorporates maximization of profit, market share growth and technology transfer as its objectives.

MD23.4 Modeling & Analysis of a 300mm Wafer Fabrication Facility

The results of an International SEMATECH project designed to use simulation to develop an understanding of factory operational issues associated with 300-mm wafer manufacturing show that the member company vision of a fully automated factory would sufficiently support the production requirements of a 300-mm factory.

MD23.5 Reuse Strategies for Discrete Simulation Software

There are many formal approaches to software re-use in the computing community. Software re-use has also long been a concern for the simulation and modeling communities. We review developments in software re-use as they apply to discrete simulation.

# Modeling Retail Operations

Session: MD24
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Modeling Retail Operations
Room:
Chair Address: University of Washington, Dept. of MS, Sch. of Bus., Box 353200, Seattle, WA 98195
Chair E-mail: kamran@u.washington.edu
Chair:
Chair E-mail:

MD24.1 Virtual vs. Brick & Mortar Retailing
• Gilles Reinhardt; RPI, Lally Sch. of Mgmt. & Tech., Troy, NY 12180-3590;
• Moren Levesque; RPI, Lally Sch. of Mgmt. & Tech., 110 Eight St., Troy, NY 12180; levesm@rpi.edu

We analyze the trade-off problem of how much a firm should supply through a brick-and-mortar network and how much it should supply through a virtual storefront. We use micro-economic models to look first at a monopolist's profit maximizing strategy and second to analyze 2 competing firms that offer a substitutable good and have access to both distribution channels...

MD24.2 no show
• Piyush Kumar; Rice University, Jones Grad. Sch., 6100 Main St., Houston, TX 77251;
• Robert A. Westbrook; Rice University, Jones Grad. Sch., 6100 Main St., Houston, TX 77251;

MD24.3 Managing Seasonal Goods Inventories with Primary & Secondary Markets
• Nicholas C. Petruzzi; University of Illinois, 350 Commerce West Bldg., 1206 South 6th St., Champaign, IL 61820;
• George E. Monahan; University of Illinois, Dept. of Business Admin., 1206 South 6th St., #350, Champaign, IL 61820; gmonahan@uiuc.edu

A retailer has a single opportunity to procure prior to a primary selling season consisting of multiple periods. Demand in each period is random, but correlated. We develop a policy to determine when the retailer should terminate the primary selling season by selling any remaining inventory on a secondary market.

Session: MD25
Date/Time: Monday 15:00-16:30
Track:
Cluster: Linear Programming & Complementarity
Room:
Chair: Tamas Terlaky
Chair E-mail:
Chair:
Chair E-mail:

MD25.1 Second-Order Cone Programming: A Survey of Applications & Solution Methods
• Robert J. Vanderbei; Princeton University, Dept. of OR & Civil Eng., Princeton, NJ 08544; rvdb@princeton.edu
• Hande Y. Benson; Princeton University, Dept. of Civil Eng. & OR, Dept. of OR & Financial Eng., Princeton, NJ 08544; hyurttan@princeton.edu

We discuss several nonlinear optimization problems that can be cast as SOCPs. We consider various formulations, some convex, some not, and study which are amenable to solution using a general-purpose interior-point solver. We compare with other NLP solvers and special codes for SOCP.

MD25.2 New Developments in Primal-Dual Interior-Point Algorithms for Convex Optimization
• Levent Tuncel; University of Waterloo, Dept. of Cominatorics & Opt., Fac. of Math, Waterloo, Ontario, N2L 3G1 , Canada; ltuncel@math.uwaterloo.ca

I will talk about generalizations of symmetric, polynomial time, primal-dual interior-point methods (and their analyses) for linear programming problems to the convex optimization problems in conic form.

MD25.3 withdrawn - author request of 10/30

# Queueing & Inventory I

Session: MD26
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: Kemal Gursoy
Chair Address: Long Island University, 1 University Plaza, H-700, Brooklyn, NY 11201
Chair E-mail: kgursoy@liu.edu,, http://phoenix.liunet.edu/~kgursoy/
Chair:
Chair E-mail:

MD26.1 Application of VARI-METRIC at the Royal Netherlands Navy

The current spare parts control method at the Royal Netherlands Navy focuses on fill rates of individual items, while it is the system availabilities that really count. We report on the potential benefits of applying an adapted VARI-METRIC method.

MD26.2 Adaptive Dynamic Programming Methods for Aging & Replenishment Problems with Applications to Inventory Mgmt. & Vehicle Dispatching
• Katerina Papadaki; Princeton University, Dept. of ORFE, Princeton, NJ 08544; papadaki@princeton.edu
• Warren B. Powell; Princeton University, Dept. of OR & Financial Eng., CASTLE Lab., Princeton, NJ 08544; powell@princeton.edu

Aging and replenishment problems arise in inventory management and vehicle dispatching problems. Standard dynamic programming methods work for simple problems but do not scale to more complex situations. We propose and study approximation methods in the context of simple problems where optimal solutions exist but have the potential of scaling to more realistic problems...

MD26.3 The Single-Site, Two-Indenture Models for Repairable Item Systems
• Zeynep M. Avsar; Bilkent University, IE Dept., Ankara, 06533 , Turkey; avsar@bilkent.edu.tr
• W. H. M. Zijm; Eindhoven University of Technology, PO Box 513, Paviljoen F18, Eindhoven, 5600 MB , The Netherlands; w.h.m.zijm@tm.tue.nl

A 2-indenture maintenance system is considered with facilities for failure detection/disassembly, component repair and assembly. A base stock policy is employed for each assembly and each component. Assuming that only one component causes an assembly failure, an approximate near-product-form steady-state distribution is proposed.

MD26.4 Multi-Item Lot-Sizing under GI/G/1 Queueing Assumptions
• Sangjin Choi; University of Calgary, Dept. of MME, Calgary, Alberta, T2N 1N4 , Canada; sjinchoi@hyowon.pusan.ac.kr
• S. T. Enns; University of Calgary, Dept. of MME, Calgary, Alberta, T2N 1N4 , Canada; enns@ucalgary.ca

Recent relationships have been developed for multi-item lot-sizing under M/M/1 and M/G/1 assumptions. These have proved valuable in understanding tradeoffs that effects system performance. However, there is still a need to develop heuristics for generalized interarrival times, which can be applied in practice. The problem and one heuristic are examined.

MD26.5 Models Dynamic Repair Allocation in Discrete Time

We consider problems of optimal allocation of repair to failed components of a system of known structure. It is assumed that the components' lifetimes and repair times are geometrically distributed with parameters that may depend on the individual component. The components are assumed to be stochastically independent, that repaired components restart as good as new and that preemption of repair is allowed...

# Stochastic Inventory Models

Session: MD27
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: Robin Roundy
Chair Address: Cornell University, Dept. of OR/IE, 216 Rhodes Hall, Ithaca, NY 14853
Chair E-mail: robin@orie.cornell.edu
Chair:
Chair E-mail:

MD27.1 Optimal Inventory Policies for Start-Up Firms
• Thomas W. Archibald; University of Edinburgh, Dept. of Bus. Studies, 50 George Square, Edinburgh Midlothian, EH8 9JY , Scotland, UK; t.archibald@ed.ac.uk,, http://www.ed.ac.uk/~twa
• Lyn C. Thomas; University of Southampton, Mgmt. School, Highfield, Southampton, Hants, SO17 1BJ , UK; l.thomas@ed.ac.uk
• John M. Betts; Monash University, Melbourne, , Australia;
• Robert B. Johnson; University of Melbourne, Melbourne, , Australia;

In this age of innovation, strategies that ensure the success of start-up firms are important. Often, the main objective of start-up firms is survival probability rather than profit. This theme is explored using Markov decision process models of inventory strategy. The problem is based on a real manufacturing start-up firm.

MD27.2 A Continuous Review Inventory Model for a Multi-Product Stochastic Demand Production System with Backorders & Budget Constraints
• Rasoul Haji; Sharif University of Technology, Dept. of IE, Azadi Ave., PO Box 11365, Tehran, , Iran; haji@www.sharif.ac.ir
• Babak Ghalebsaz Jeddi; Pars Pad Co., PO Box 13145-1855, Tehran, , Iran; ghalebsaz@apadana.com

In the literature of production and inventory control, there has been presented a continuous review inventory model for a single product stochastic demand unconstrained system where backorders are allowed. We consider the same model and we extend it for the case of multi-product inventory system with budgetary constraints.

MD27.3 Base-Stock Policies for Lost-Sales Problems

We consider base-stock policies for a single installation facing stochastic demand. Demand that is not met is lost. The proof of the convexity of the objective function is presented.

# Threats to Life & Limb

Session: MD28
Date/Time: Monday 15:00-16:30
Sponsor: Public Programs & Processes Section
Track:
Cluster:
Room:
Chair: Arnold I. Barnett
Chair Address: MIT, Sloan Sch. of Mgmt., Cambridge, MA 02139-4307
Chair E-mail: abarnett@mit.edu
Chair:
Chair E-mail:

MD28.1 Too Many Hospital Beds?
• Linda V. Green; Columbia University, Grad. School of Bus., New York, NY 10027; lvg1@columbia.edu

Excess hospital capacity has been cited as a major factor contributing to the high cost of health care. As a result, significant reductions in the number of hospital beds continue to be made. Are there too many beds? What's the impact of bed reductions on patient care? We focus on these questions and related issues.

MD28.2 Factors in the US Homicide Drop
• Alfred Blumstein; Carnegie Mellon University, Heinz School of PP&M, NCOVR, 5000 Forbes Ave., Pittsburgh, PA 15213-3890; ab0q@andrew.cmu.edu

US homicide rates dropped more than 40% between the 1991 peak and 1999. We identify key factors contributing to the drop, including removing handguns from kids, tightening gun supply, reduction of new entrants to crack markets and the robust economy. We also examine the role of the growth of incarceration.

MD28.3 Allocating HIV Prevention Resources
• Edward Kaplan; Yale University, Sch. of Organization & Mgmt., Sch. of Med., New Haven, CT 06520-8200; edward.kaplan@yale.edu

The Institute of Medicine's Committee on HIV Prevention Strategies was charged with proposing a new national strategy for HIV prevention. Focusing on the allocation of HIV prevention resources, the committee estimated the value of wiser resource targeting. In summarizing the committee's report, we highlights OR ideas and contributions.

MD28.4 Runway Collisions: Crying Wolf?
• Arnold I. Barnett; MIT, Sloan Sch. of Mgmt., Cambridge, MA 02139-4307; abarnett@mit.edu
• Gary Paull; MCR Federal, Burlington, MA;

Both the National Transportation Safety Board and the Federal Aviation Administration have declared the danger of runway collisions as the number-one threat to US aviation safety. Analyzing various data sets, we assess whether this concern is justified. (Hint: It is.)

# Price Volatility & Probabilistic Methods in the Energy Market

Session: MD29
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: Shi-Jie Deng
Chair Address: Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205
Chair E-mail: deng@isye.gatech.edu
Chair:
Chair E-mail:

MD29.1 Current Challenges in Modeling Power Price Volatility

Developments in US power markets have created several unique challenges for energy industry economists. We discuss the major factors underlying the exceptionally high volatility of electricity prices. We feel that some of them may reflect flaws in power-pools design and incomplete transition to fully deregulated markets in generation and transmission.

MD29.2 Jump-Diffusion Models in Pricing Energy Options
• John R. Birge; Northwestern University, McCormick Sch. of Engineering, 2145 Sheridan Rd., Evanston, IL 60208-3100; jrbirge@nwu.edu
• S. G. Kou; Columbia University, Dept. of IE/OR, Mudd Bldg., New York, NY 10027; kou@ieor.columbia.edu

We will present 2 types of models for pricing energy options with jump-diffusion processes. Each may relate to different types of markets depending on the characteristics of the jump process. One appears more consistent for gas pricing while the other is more suitable for electricity pricing.

MD29.3 California Electricity Prices: A Spectral Analysis of Changing Price Patterns as a Result of Price Caps
• Rajesh Rajaraman; ;

We recently applied spectral frequency domain methods to help with the characterization of price volatility process to be able to separate out predictable components of electricity prices from volatility. The results were applied to PJM prices. The methodology is applied to a sliding window of California zonal prices to try to understand the effect of price caps on predictability and volatility of prices.

MD29.4 An Alternative Model for Electricity Spot Prices
• Shi-Jie Deng; Georgia Institute of Technology, Sch. of ISyE, Atlanta, GA 30332-0205; deng@isye.gatech.edu

Unique characteristics of the electricity spot price make its modeling an extremely challenging task. We present an approach for modeling marginal distributions of power prices utilizing quantile functions. Such an approach is very appealing for purpose of empirical parameter estimation since simulations based on the resulting distributions can be easily performed.

MD29.5 A Probabilistic Method for Estimating Future Growth of Oil & Gas Reserves in the US
• Robert A. Crovelli; US Geological Survey, Denver Federal Ctr., PO Box 25046, MS 939, Denver, CO 80225; crovelli@usgs.gov
• James W. Schmoker; US Geological Survey, Denver Federal Ctr., Box 25046, MS 939, Denver, CO 80225;

The Probabilistic Reserve Growth Spreadsheet system was developed from analytic probabilistic methodology to calculate estimates of future growth of crude oil and natural gas reserves in the US, where reserve growth is strongly positive and is a major component of remaining US oil and natural-gas resources.

# Data Mining Applications

Session: MD30
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: W. Dwight Calkins
Chair Address: IBM, MD 28-04-4003, 1503 LFB Freeway, Dallas, TX 75234
Chair E-mail: dcalkin@us.ibm.com
Chair:
Chair E-mail:

MD30.1 Using the SAS Enterprise Miner to Investigate Retail Consumer Segmentations using Data Obtained from the Bureau of Labor Statistics
• W. Dwight Calkins; IBM, MD 28-04-4003, 1503 LFB Freeway, Dallas, TX 75234; dcalkin@us.ibm.com

The Consumer Expenditure Survey obtained from the USA Department of Labor Statistics can be datamined using the SAS Enterprise Miner in order to classify consumer retail behavior in the US. After a particular retail expenditure category is selected for a given study, i.e., personal computer purchases, the resulting segmentation then defines the consumers in terms of a variety of demographic variables...

MD30.2 Mass Customization of Direct Marketing Materials using Data Mining Technology

An integrated use of data mining, OR and recent advances in printing technology are described for use in mass customization of direct mail and web-based marketing materials.

MD30.3 The Integration of Discovery- & Non-Discovery-Based Data Mining with Spatial Analysis Techniques to Create Solutions
• Frederick D. Busche; IBM, 3116 Lake Highlands Dr., Highland Village, TX 75077; fbusche@us.ibm.com

One aspect of data can be an entity and the attributes that represent that entity. The second is the rows and rows of this entity data. The third aspect of these data is the spatial relationship or location of the entity. We discuss the integration of processes that allow for the evaluation of these data.

MD30.4 Remapping Customer Relationship Management to the E-Business Environment

Our value proposition is to maximize customer portfolio results by providing applications and services that drive the profitability of your e-business customer relationships through the adaptation of relevant messages and synchronization across any customer touchpoints. This is accomplished using advanced data warehousing methods, deep analytical pattern recognition and mathematical optimization techniques.

# Applying Optimization Technologies: Lectures of GMU Students Honoring Carl M. Harris

Session: MD31
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: Karla Hoffman
Chair Address: George Mason University, Systems Eng. & OR Dept., 4400 University Dr., MS 4A6, Fairfax, VA 22030
Chair E-mail: khoffman@gmu.edu,, http://iris.gmu.edu/~khoffman
Chair:
Chair E-mail:

MD31.1 A Nonlinear Rescaling Principle for Large-Scale Constrained Optimization: A Primal-Dual Approach
• Igor Griva; George Mason University, Dept. of SEOR, 4400 University Dr., Fairfax, VA 22030;
• Roman A. Polyak; George Mason University, Dept. of SEOR, 4400 University Dr., Fairfax, VA 22030-4444;

The nonlinear rescaling principle consists of rescaling constraints into an equivalent constraint set and using the Lagrangian for the equivalent problem. Newton's method is performed on the primal-dual system of equations consisting of both the optimality criteria equations and the Lagrange multipliers update. Convergence properties, numerical stability and numerical results on large-scale problems will be presented.

MD31.2 Modeling to Optimize Restoration Tracking & Investments
• Linda Coblentz; Center for Army Analysis, Ft. Belvoir, VA;

The Center for Army Analysis developed an integer programming model to provide alternative schedules for environmental restoration projects, based on different objective functions. Restoration projects are broken down into phases that must occur sequentially and with specified phase lengths. To date, 2 objective functions have been used.

MD31.3 Mobile Optimization for the Scheduling of Transit Operations
• Martin T. Durbin; Decisive Analytics Corp., 1235 Jefferson Davis Highway, Ste. 400, Arlington, VA;
• Karla Hoffman; George Mason University, Systems Eng. & OR Dept., 4400 University Dr., MS 4A6, Fairfax, VA 22030; khoffman@gmu.edu,, http://iris.gmu.edu/~khoffman

MOST optimizes the deployment and scheduling of a fleet of delivery vehicles in real-time. The software enables driver assignments and delivery schedules to be revised continuously, taking into account new high-priority orders, traffic and weather conditions and other unforeseen events. An intuitive interface allows the user to easily modify the current schedule while showing impacts on overall order fulfillment.

# Telecommunication Systems III

Session: MD32
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: John G. Klincewicz
Chair Address: AT&T Labs., 200 Laurel Ave., Rm. D5-3C06, Middletown, NJ 07748
Chair E-mail: klincewicz@att.com
Chair:
Chair E-mail:

MD32.1 Routing & Restoration in Optical Networks
• Iraj Saniee; Bell-Labs, Lucent Technologies, 600 Mountain Ave., Rm. 2C-300, Murray Hill, NJ 07974; iis@bell-labs.com
• Gang Liu; Bell-Labs, Lucent Technologies, 600 Mountain Ave., Rm. 2C-300, Murray Hill, NJ 07974;
• Eric Bouillet; Bell-Labs, Lucent Technologies, 600 Mountain Ave., Rm. 2C-300, Murray Hill, NJ 07974;

We describe multiple schemes for restoration of wave-length routes in optical networks. We define and quantify critical metrics associated with these schemes such as cost, reconfigurability, port and wavelength efficiency and restoration times. We conclude with a description of 4 restoration classes, calculate their metrics and outline their associated routing algorithms on a large network.

MD32.2 Robust Location Problems on Trees

We consider different aspects of robust 1-median problems on trees with uncertain or dynamically changing edge lengths and vertex weights which can also have negative weights. The dynamic nature of a parameter is modeled by a linear function of time. The uncertainty is modeled by given intervals, in which each parameter can take a value randomly...

MD32.3 Delay Allocation in a Packet Network with Quality of Service
• John G. Klincewicz; AT&T Labs., 200 Laurel Ave., Rm. D5-3C06, Middletown, NJ 07748; klincewicz@att.com

Different classes of traffic within a corporate intranet must meet different delay constraints. For a given class, we describe a heuristic for allocating a maximum allowable delay to each link in a network, such that for all node pairs, the end-to-end delays meet a pre-specified constraint. Motivation for this heuristic, in terms of an exact solution for a simple version of the problem, is discussed.

MD32.4 The Value of Combinatorial Auctions in Telecom Trading
• Lakshman P. Sinha; Telcordia Technologies, Dept. of Applied Research, 445 South St., Rm. 1A328B, Morristown, NJ 07960; lsinha@research.telcordia.com
• Siddhartha Dalal; Telcordia Technologies, Dept. of Applied Research, 445 South St., Rm. 1J304R, Morristown, NJ 07960; sid@research.telcordia.com
• Robert E. Hausman; Telcordia Technologies, Dept. of Applied Research, 445 South St., Rm. 1C349B, Morristown, NJ 07960; rhausman@telcordia.com
• Tracy Mullen; Telcordia Technologies, Dept. of Applied Research, 445 South St., Rm. 1D312B, Morristown, NJ 07960; mullen@research.telcordia.com

Telecom bandwidth and minutes are rapidly becoming commodities and are now traded as such on many exchanges e.g., Arbinet, RateXchange, Band-X and more. To the best of our knowledge, telecom trading is currently done for one item at a time. We present initial results on the value of using combinatorial auctions in telecom trading.

# Flowshop Scheduling

Session: MD33
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: Jerzy Kamburowski
Chair Address: University of Toledo, Coll. of Bus. Admin., Toledo, OH 43606
Chair E-mail: jkambur@utnet.utoledo.edu
Chair:
Chair E-mail:

MD33.1 A New Heuristic for an m-Machine Flowshop with the Total Completion Time Criterion
• Ali Allahverdi; Kuwait University, Dept. of MIE, PO Box 5969, Coll. of Eng. & Petroleum, Safat, 13060 , Kuwait; allahverdi@kuc01.kuinv.edu.kw
• Tariq Aldowaisan; Kuwait University, Dept. of MIE, PO Box 5969, Coll. of Eng. & Petroleum, Safat, 13060 , Kuwait; tariq@kuc01.kuinv.edu.kw

We consider an m-machine flowshop with the objective of minimizing total completion times of all the jobs. We compare the 2 recent independently developed heuristics and propose a new heuristic. Computational analysis shows that the proposed heuristic is superior to the most recent ones.

MD33.2 A Genetic Algorithm-Tabu Search Hybrid Heuristic for Permutation Flow-Shop Scheduling

We introduce a hybrid genetic algorithm-tabu search heuristic for the minimal makespan flow shop sequencing problem. In order to evaluate the effectiveness of the hybridization, we compare the hybrid metaheuristic with both pure GA and tabu search heuristics. Results from computational experience are discussed.

MD33.3 Heuristics for Hybrid Flow Shop Scheduling

We deal with the k-stage hybrid flow shop scheduling problem where, at any stage, there are 2 identical machines. We introduce 2 heuristic procedures with the objective of minimizing makespan. Computational experience was carried out in order to evaluate the performance of the heuristics. The main results are presented.

MD33.4 Efficiently Solvable Special Cases of the Three-Machine Flowshop Problem

The well-known NP-hard problem of finding the job sequence that minimizes the makespan in a 3-machine flow shop is reconsidered. We define 2 new dominance relations on the job processing times and present corresponding heuristics that yield the optimal sequence for all efficiently solvable special cases identified so far.

# Multicriteria Decision Making II

Session: MD34
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: Noel Artiles-Leon
Chair Address: University of Puerto Rico, IE Dept., PO Box 9043, Mayaguez, PR 00681-9043
Chair E-mail: n_artiles@rumac.uprm.edu
Chair:
Chair E-mail:

MD34.1 Balancing Conflicting Objectives in the News Vendor Problem
• Mahmut Parlar; National University of Singapore, Dept. of Dec. Sci., Singapore, 119260 , Singapore; parlar@mcmaster.ca
• Z. Kevin Weng; University of Wisconsin, Sch. of Bus., Madison, WI 53706;

We consider the news vendor problem as a bi-criteria optimization problem with 2 objectives: maximize the expected profit and maximize the probability of exceeding the expected profit. We use compromise programming to find the solution on the efficient frontier that has the shortest distance from the ideal solution.

MD34.2 The Compromise Hypersphere for Multiobjective Linear Programming
• Saul I. Gass; University of Maryland, 8809 Maxwell Dr., Potomac, MD 20854-3123; sgass@rhsmith.umd.edu
• Pallabi Guha Roy; University of Maryland, Dept. of Math., College Park, MD 20742;

We propose a method for ranking efficient solutions to a multiobjective LP problem. The solutions, as points in objective space, are enclosed within an annulus of minimum width, where the width is determined by a hypersphere that minimizes the maximum deviation of the points from the surface of the hypersphere.

MD34.3 Multi-Objective Optimization of the Low Temperature NOx Control Process
• Yan Fu; Ford Motor Co., Ford Research Lab., MD 2115, Rm. 2629, SRL, Dearborn, MI 48121; yfu4@ford.com
• Urmila Diwekar; Carnegie Mellon University, Baker Hall 129, Dept. of EPP, 5000 Forbes Ave., Pittsburgh, PA 15213; ud01@andrew.cmu.edu

Designing the best available control technology for NOx removal is a multiobjective optimization problem. A new and efficient multi-objective optimization algorithm based on a novel sampling technique was developed and successfully applied to obtain minimum cost and minimum pollution emissions designs for this large-scale real-world problem.

MD34.4 Bootstrapping-Based Confidence Regions for Multiple-Response Problems
• Noel Artiles-Leon; University of Puerto Rico, IE Dept., PO Box 9043, Mayaguez, PR 00681-9043; n_artiles@rumac.uprm.edu
• Narcisa Meza; University of Puerto Rico, IE Dept., PO Box 9043, Mayaguez, PR 00681-9043;

We present a rational and practical approach to the problem of simultaneously optimizing several response variables and constructing a confidence region for the optimal settings. The methodology uses a desirability function to find the optimal settings and bootstrapping to construct their confidence regions.

# Information Systems V

Session: MD35
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: Kenny Kwong
Chair Address: PricewaterhouseCoopers LLP, 13201 South Wakial Loop 2082, Phoenix, AZ 85044
Chair E-mail: kenny.kwong@us.pwcglobal.com,, http://www.pwcglobal.com
Chair:
Chair E-mail:

MD35.1 'Different Slopes for Different Folks': Identifying the Source of Performance Variation following Implementation of Enterprise-Level IT

We ask from what level within the firm does variation in the rate of performance improvement arise following IT implementation. Understanding the source of this variation is crucial if research is to further reveal ways to accelerate it. Findings from a study of ERP implementation in 24 sites suggest a process-level approach to understanding enterprise-level information technology implementations.

• Larry J. LeBlanc; Vanderbilt University, Owen Grad. Sch. of Mgmt., Nashville, TN 37203-9865; larry.leblanc@owen.vanderbilt.edu
• James A. Hill; Vanderbilt University, Owen Grad. Sch. of Bus., Nashville, TN 37203;
• David M. Dilts; Vanderbilt University, Owen Grad. Sch. of Mgmt., Nashville, TN 37203;

We model the effects of sharing information in 3-level supply chains. Using simulation, we quantify the benefits of sharing information in terms of faster order fulfillment. When customers change their orders, information sharing can be harmful. Specifically, increased inventories result when vendors produce in anticipation of orders that are later cancelled.

MD35.3 Language & Electronic Commerce
• Jeffrey M. Keisler; University of Massachusetts, 100 Morrissey Blvd., M/5-230, Boston, MA 02125; jeff.keisler@umb.com

Transactions in e-commerce must meet each side's goals. Offers and requirements both consist of enactable first order logical statements. Using Craig interpolants, we examine how the set of achievable deals depends on customer and vendor languages and explore theoretical and real world implications of formal language as a strategic asset.

MD35.4 Leveraging E-Procurement in Supply Chain Management/Master Planning

We discuss how emerging technologies in e-procurement augment supply chain and master planning tools and expedite the planning cycle. We discuss prevalent architecture and best practices for integrating master planning tools with third party e-procurement tools. One of the most observable benefits of this integration is the real-time availability of data using the simplest communication platform - the Web...

# Operations Management III

Session: MD36
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: Ishpal S. Rekhi
Chair E-mail: irekhi@csusm.edu
Chair:
Chair E-mail:

MD36.1 An Order-up-to-Level Updating Mechanism for Inventory Systems with Unobserved Lost Sales

We consider an inventory management problem in the retail industry with unobserved lost sales. The retailer has access to point-of-sale (POS) data and uses a fixed review period, order-up-to level system to control the inventory of a particular product to achieve a pre-specified service level. However, due to the unobserved lost sales nature of the problem, the retailer cannot exactly measure the current service level...

MD36.2 The Effect of Integration of Manufacturing Practices on Performance
• Kristy O. Cua; University of Minnesota, 321 19th Ave. South, Ste. 3-150, Minneapolis, MN 55455; kcua@csom.umn.edu
• Roger Schroeder; University of Minnesota, Carlson Sch. of Mgmt., 321 19th Ave. South, Ste. 3150, Minneapolis, MN 55455; rschroeder@csom.umn.edu
• Kathleen E. McKone; Babson College, Babson Hall 211A, Babson Park, MA 02457; kmckone@babson.edu

We investigate the effect of implementation of manufacturing practices and find that co-alignment of practices is positively related to performance. There are different configurations of co-alignment of practices depending on the performance measure being emphasized but good performance in several dimensions can be achieved simultaneously.

MD36.3 Knowledge Acquisition & Transfer in Service Settings: Customer Outrage in Airlines
• Michael A. Lapre; Boston University, Sch. of Mgmt., 595 Commonwealth Ave., Boston, MA 02215; mlapre@bu.edu,, http://people.bu.edu/mlapre/
• Nikos Tsikriktsis; London Business School, Sussex Place, Regent's Park, London, NW1 4SA , UK;

We analyze customer complaint learning-curves in the US domestic airline industry using data for all major airlines available since the Department of Transportation introduced service quality indicators. Geographic specialist airlines reduced complaints via learning-by-doing, whereas generalist airlines did not. Experience from other airlines employing similar strategies also reduced complaints.

MD36.4 Dynamic Stock Allocation & its Implication on Delayed Product Differentiation
• Francis de Vericourt; Bouygues Telecom, RED Europa L 30 av de l'Europe, Velizy, 78944 , France; fdeveric@bouyguestelecom.fr
• Fikri Karaesmen; ECP, Lab. Product. et Logistique, Grande Voie des Vignes, Chatenay-Malabry, 92295 , France; fikri@pl.ecp.fr
• Yves P. Dallery; ECP, Lab. Product. et Logistique, Grande Voie des Vignes, Chatenay-Malabry, 92295 , France; dallery@pl.ecp.fr

We study dynamic stock allocation problems with different types of demand for a single item. We present results on the structure of optimal dynamic allocation policies and propose a framework which enables a characterization of the benefits of delayed product differentiation in terms of inventory related costs.

MD36.5 Allocation of Buffer in Unbalanced Parallel Assembly Systems
• Ishpal S. Rekhi; California State University, Coll. of Business Admin., San Marcos, CA 92096-0001; irekhi@csusm.edu

We consider a parallel assembly system where 2 or more units are processed at different workstations and the final product is assembled at an assembly station. We consider processing distributions that are not identical. We study the optimal location and allocation of buffer in this system and the factors that dictate the optimal placement of buffers.

# Linear Programming & Applications

Session: MD37
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: Harold J. Schleef
Chair Address: Lewis & Clark College, 0615 SW Palatine Hill Rd., Portland, OR 97219
Chair E-mail: schleef@lclark.edu
Chair:
Chair E-mail:

MD37.1 Stripping Coupons with Linear Programming
• Harry Zheng; University of Southampton, Sch. of Mathematics, Southampton, Hants, SO17 1BJ , UK; h.zheng@ed.ac.uk
• Lyn C. Thomas; University of Southampton, Mgmt. School, Highfield, Southampton, Hants, SO17 1BJ , UK; l.thomas@ed.ac.uk
• David E. Allen; Edith Cowan University, Sch. of Bus. Econ. & Finance, Joondalup, WA 6027 , Australia; d.allen@cowan.edu.au

The yield curve of interest rates is obtained from bond prices. Most bonds have coupons which means there is a stream of payments which need to be translated into the value of a unit payment at each time point. Traditionally, this has been done by statistical methods including bootstrapping and fitting spline curve structures but these can lead to mispricing of risky bonds...

MD37.2 Large-Sample Analysis of Linear Programming under Random Data Perturbations

We present results concerning laws of large numbers, central limit theorems and general convergence of optimal vector solutions and optimal objective values of LPs in the context of random data perturbations.

MD37.3 Proximity Functions for Linear & Semidefinite Optimization

Interior-point methods can use small or large updates of the barrier parameter. Contrary to theoretical results, large updates are more efficient than small updates, in practice. We discuss the classic approach and show why and how better complexity results can be obtained by using a so-called Upsilon-concordant proximity function.

MD37.4 A Priori Error Bounds for the Aggregated Generalized Transportation Problem
• Igor S. Litvinchev; Computing Center Russian Academy of Sciences, Arce 5169, Col. Valle Verde, Monterrey Nuevo Leon, 64360 , Mexico;
• Socorro Rangel; UNESP, CP 136, SJ Rio Preto, SP, 15054-000 , Brazil; socorro@nimitz.dcce.ibilce.unesp.br
• Oscar Chacon; UANL, Arce 5169, Col. Valle Verde, Monterrey Nuevo Leon, 64360 , Mexico; ochacon@uanl.mx

In contrast to the CTP, a priori error bounds have not been developed yet for the generalized one. We propose the a priori bounds for the customers aggregation in the GTP. Numerical experiments show a strong and statistically significant correlation between the actual aggregation error and the new a priori bound. This gives the analyst a key to compare various types of aggregation before an aggregated problem was solved.

MD37.5 Structuring IRA Settlements with Roth IRAs
• Harold J. Schleef; Lewis & Clark College, 0615 SW Palatine Hill Rd., Portland, OR 97219; schleef@lclark.edu

Settlements related to incorrect distributions of traditional IRAs may have substantial tax consequences. The tax characteristics of the Roth IRA make it a useful instrument for reducing the payment required of the IRA issuer. Several examples using LP are presented.

# Semiconductor Industry

Session: MD38
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: Michael Pangburn
Chair Address: Pennsylvania State University, 315 Beam Bldg., MSIS Dept., Smeal Coll. of Bus., University Park, PA 16802-1913
Chair E-mail: mikepangburn@psu.edu,, msp5@psu.edu
Chair:
Chair E-mail:

MD38.1 Anoxide Materials for Electronic Techniques
• Andrey G. Petrov; , 3 Basseynaya St., Ste. 11, Kiev, 01004 , Ukraine; lgi@lgi.ru.kiev.ua

We cover the aspects of chemistry, technology and physics of the complex chalcohenyd and chalcohalohenyd compounds. Theoretical basis of the searching and forecasting new compounds with prior set properties is discussed. Synthesis & growing monocrystals, their basic physical and chemical characteristics and structure. A theory of the new ferroelectrics-semiconductors, gyrotropous materials with pseudoisotropic point and other composites is presented.

MD38.2 Experiments with the Optimal Capacity Expansion Technique FIFEX

We apply a technique to determine optimal machine, floor and shell expansion times in semiconductor fabs in the presence of uncertain product demands. We use real life demand, capacity data and explicitly account for the resolution of demand uncertainty. We run 4 sets of experiments to determine the effects of rounding expansion times, cost savings over some heuristics...

MD38.3 An Operation Theory Approach to the Costing of Customer Satisfaction

Customer satisfaction is essential in the semiconductor business, but its future value and potential costs are difficult to estimate. We present a quantitative approach combining customer service with operations planning to estimate the present costs of customer satisfaction policies and to provide guidelines to the long-term value expected from such policies.

MD38.4 Capacity Planning & Intertemporal Pricing for Semiconductor Production
• Michael Pangburn; Pennsylvania State University, 315 Beam Bldg., MSIS Dept., Smeal Coll. of Bus., University Park, PA 16802-1913; mikepangburn@psu.edu,, msp5@psu.edu
• Shankar Sundaresan; Pennsylvania State University, 316 Beam Bldg., University Park, PA 16802; shankar-s@psu.edu

Semiconductor producers make significant and long-term capital expenditures in manufacturing technology. When setting manufacturing capacity, firms must assess how prices will change over time, as well as the impact of subsequent product introductions. By developing economic models that significantly extend previous manufacturing studies, we provide insights into this capacity-setting problem.

# New Product Development II

Session: MD39
Date/Time: Monday 15:00-16:30
Type: Contributed
Track:
Cluster:
Room:
Chair: Pirooz Vakili
Chair Address: Boston University, Dept. of Mfg. Eng., 15 Saint Mary's St., Boston, MA 02215
Chair E-mail: vakili@bu.edu
Chair:
Chair E-mail:

MD39.1 A Decision Model for Managing New Product Development Risks
• Sharon A. Johnson; Worcester Polytech Institute, Dept. of Mgmt., 100 Institute Rd., Worcester, MA 01609; sharon@wpi.edu
• Gary M. Searle; Worcester Polytech Institute, Dept. of Mgmt., 100 Institute Rd., Worcester, MA 01609;

A decision model for managing project risk in new product development projects was developed based on project and organizational objectives. Responses about risks are elicted from amanagement team and combined with progress data to create risk scorecards. Two case studies were used to develop, refine and test the model.

MD39.2 Evaluation of the Dynamic Screen Builder
• Michael F. Johnson; USAA Safety & Environmental Affairs, 9800 Fredricksburg Rd., San Antonio, TX 78228-0062; michael.johnson@usaa.com
• Mary Garcia; USAA Safety & Environmental Affairs, 9800 Fredricksburg Rd., San Antonio, TX 78228-0062; mary.garcia1@usaa.com
• Hank Austin; USAA Safety & Environmental Affairs, 9800 Fredricksburg Rd., San Antonio, TX 78228-0062; hank.austin@usaa.com
• Rafael Moras; St. Mary's University, One Camino Santa Maria, San Antonio, TX 78228-8534; moras@stmarytx.edu

The implications of software design relative to ergonomics are rarely investigated. We discuss the ramifications resulting from the introduction of a data entry package at a major insurance company. The use of surveys, statistics and ergonomic principles to evaluate the effects of this system on claims representatives is reported.

MD39.3 Dynamic Management of New Product Development Projects & Portfolios: A Modeling Framework
• Pirooz Vakili; Boston University, Dept. of Mfg. Eng., 15 Saint Mary's St., Boston, MA 02215; vakili@bu.edu
• Yanfeng Wang; Boston University, 15 Saint Mary's St., Boston, MA 02215; wyanfeng@bu.edu
• Anil Khurana; Boston University, Sch. of Mgmt., 595 Commonwealth Ave., Boston, MA 02215; akhurana@bu.edu

We describe a modeling framework for the management of new product development projects and portfolios. Tradeoffs when setting interim project performance targets and uncertainty in achieving them are explicitly modeled and budget allocation decisions in this context are considered. Some structural properties of optimal policies are derived and their implications in specific examples are illustrated.

# Structural Models of Retail Competition

Session: MD40
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: Structural Models of Retail Competition
Room:
Chair: Sachin Gupta
Chair Address: Cornell University, Johnson Grad. Sch. of Mgmt., Sage Hall, Ithaca, NY 14853
Chair E-mail: s-gupta7@nwu.edu,, sg248@cornell.edu
Chair: Jean-Pierre Dube
Chair Address: University of Chicago, Grad. School of Bus., 1101 East 58th St., Chicago, IL 60637
Chair E-mail: jdube@gsbux1.uchicago.edu

MD40.1 Price Competition in Industries with Geographic Differentiation: The Case of Fast Food
• Raphael Thomadsen; Columbia University, Grad. School of Bus., 3022 Broadway, 613 Uris Hall, New York, NY 10027; raphael@stanford.edu

We have collected a data set of prices and locations of fast food restaurants in Santa Clara County, California. We fit this data to a model of supply and demand for fast food by estimating an indirect utility function for consumers and marginal costs for firms using a generalized method of moments estimator...

MD40.2 Endogeneity & Heterogeneity in Discrete Choice Models with Panel

We consider the existence of general correlations of error terms and propose a robust way to test whether the endogeneity problem is significant in a given data set under weak assumptions on the error structure that allow for heterogeneity and common shocks. The test is a test of the differences in GMM coefficient estimates of a model with and without instrumenting for right hand side variables...

MD40.3 Price Discrimination in Gasoline Markets: A Theoretical & Empirical Analysis
• Ganesh Iyer; Washington University, Olin Sch. of Bus., CB 1133, One Brookings Dr., St. Louis, MO 63130; iyer@olin.wustl.edu
• P. B. Seetharaman; Washington University, Olin Sch. of Business, 1 Brookings Dr., St. Louis, MO; seethu@olin.wustl.edu

We investigate price-discrimination in retail gasoline markets. We provide a theoretical analysis of price discrimination in this market in which a gas station's decision to price discriminate and the extent of price discrimination are endogenously determined by demographic variables characterizing its local market, the brand name and other station characteristics. We also do empirical testing.

MD40.4 Heterogeneity & Target Marketing using Aggregate Retail Data: A Structural Approach
• Jean-Pierre Dube; University of Chicago, Grad. School of Bus., 1101 East 58th St., Chicago, IL 60637; jdube@gsbux1.uchicago.edu
• David Besanko; Northwestern University, Kellogg Grad. Sch. of Mgmt., 2001 Sheridan Rd., Evanston, IL 60208; d-besanko@nwu.edu
• Sachin Gupta; Cornell University, Johnson Grad. Sch. of Mgmt., Sage Hall, Ithaca, NY 14853; s-gupta7@nwu.edu,, sg248@cornell.edu

We develop a heterogeneous logit model of consumer demand jointly with a structural model of pricing. We include a vertical channel consisting of competing manufacturers and a monopolistic retailer. We represent heterogeneity using the popular finite mixture approach, wherein consumers belong to a small number of latent classes that differ in their preferences and responsiveness to the marketing mix.

MD40.5 Category Pricing

Category management is a critical function of retailers. Advances in the adoption of optical scanners and better use of information and technology give retailers detailed data that can be used towards this purpose. The recent proliferation of new product introductions by manufacturers has increased the importance of retail category management. One key aspect of retail category management is pricing the brands within a product category...

# OR Applications

Session: MD41
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: Kaan Katircioglu
Chair Address: IBM Integrated Supply Chain, TJ Watson Research Ctr., Rte. 134, Yorktown Heights, NY 10598
Chair E-mail: kaan@us.ibm.com
Chair:
Chair E-mail:

MD41.1 Drivers of Performance in the Semiconductor Industry
• Kaan Katircioglu; IBM Integrated Supply Chain, TJ Watson Research Ctr., Rte. 134, Yorktown Heights, NY 10598; kaan@us.ibm.com
• William Grey; IBM Integrated Supply Chain, TJ Watson Research Ctr., Rte. 134, Yorktown Heights, NY 10598; williamgrey@us.ibm.com
• Robert J. Baseman; IBM Integrated Supply Chain, TJ Watson Research Ctr., Rte. 134, Yorktown Heights, NY 10598; robertbaseman@us.ibm.com
• Arnold B. Maltz; Arizona State University, PO Box 874706, Tempe, AZ 85287-4076; arnie.maltz@asu.edu
• William Grenoble; Pennsylvania State University; wlg2@psu.edu
• Dale Rogers; University of Nevada; mickey@unr.edu

We investigate the factors that impact performance in semiconductor supply chains through regression analysis. This study is based on a set of detailed data collected in major semiconductor companies in the US. Our analysis reveals interesting results about the drivers of performance and their impact.

MD41.2 A Successful Implementation of a Stochastic Inventory Model in the Semiconductor Industry

We have successfully implemented a business process for managing inventory at Xilinx. It was developed around IBM's stochastic inventory system: the Asset Management Tool, recipient of the 1999 Edelman Award for Achievement in OR. This work required development of cost and flow management models uniquely adapted to the semiconductor industry.

MD41.3 Lessons for OR/MS Practice from the History of Edelman Finalist Corporations
• Peter C. Bell; University of Western Ontario, Ivey Sch. of Bus., 1151 Richmond St., London, Ontario, N6A 3K7 , Canada; pbell@ivey.uwo.ca

MS/OR groups that are finalists in the Edelman competition grab the spotlight for a moment, but what happens after that? We have contacted all the private sector Edelman finalists from 1989-1996 to find out what happened to OR/MS, and specifically the Edelman applications, in the 5 years following the competition. The results reveal important lessons for the successful practice of MS/OR.

# Statistical Quality Control with Complex-Structured Data & Processes

Session: MD42
Date/Time: Monday 15:00-16:30
Sponsor: Quality, Statistics & Reliability Section
Track:
Cluster: Reliability & Quality Control
Room:
Chair: Fugee Tsung
Chair Address: HKUST, Dept. of IE/EM, Clear Water Bay, Kowloon, , Hong Kong
Chair E-mail: season@ust.hk
Chair: Daniel W. Apley
Chair Address: Texas A&M University, Dept. of IE, College Station, TX 77843-3131
Chair E-mail: apley@tamu.edu

MD42.1 Multistage Process Monitoring & Diagnosis
• Fugee Tsung; HKUST, Dept. of IE/EM, Clear Water Bay, Kowloon, , Hong Kong; season@ust.hk

As manufacturing quality has become a decisive factor in global market competition, quality control techniques such as statistical process control and automatic process control are becoming popular in industries. With advances in information, sensing and data capture technology, large volumes of data are being routinely collected and shared over multiple-stage processes...

MD42.2 Analysis of Diagnosability of Multistage Manufacturing Processes using a State Space Approach
• Yu Ding; University of Michigan, Dept. of IOE, Ann Arbor, MI 48109;
• Darek Ceglarek; University of Wisconsin, IE Dept. 266E Mech. Eng. Bldg., 1513 University Ave., Madison, WI 53706-1572; darek@engr.wisc.edu
• Jan Shi; University of Michigan, Dept. of IOE, 1724 IOE Bldg., Ann Arbor, MI 48019-2117; shihang@umich.edu

Based on the state space model of MMPs, the diagnosability of MMPs is studied following a similar concept of observability in control theory, by defining diagnosability matrix and index and deriving their expressions in terms of process/product design. Case studies are presented to validate the proposed methodology.

MD42.3 Robust Minimum Variance Control with Model Uncertainty
• Jeongbae Kim; Texas A&M University, Dept. of IE, College Station, TX 77843-3131;
• Daniel W. Apley; Texas A&M University, Dept. of IE, College Station, TX 77843-3131; apley@tamu.edu

Given the nature of industrial processes and the manner in which their models are estimated, model uncertainty is unavoidable. It is well known that the performance of minimum variance control strategies deteriorates in the presence of model uncertainty, to the point that variance may actually be increased. Existing robust control strategies incorporate deterministic measures of model uncertainty...

MD42.4 An Optimal Filter Design Approach to Statistical Process Control for Autocorrelated Processes
• Chang-Ho Chin; Texas A&M University, Dept. of IE, College Station, TX 77843-3131;
• Daniel W. Apley; Texas A&M University, Dept. of IE, College Station, TX 77843-3131; apley@tamu.edu

EWMA and Shewhart charts, on either the original data or time-series residuals, are 2 of the most widely used methods for SPC of auto-correlated processes. The charted statistic can be viewed as the output of a linear filtering operation on the original data. We generalize this concept, with the goal of designing an 'optimal' general linear filter, the output of which is the SPC statistic to be charted...

# Best Dissertation Award Winners 2000

Session: MD43
Date/Time: Monday 15:00-16:30
Track:
Cluster:
Room:
Chair: Jeffrey K. Liker
Chair Address: University of Michigan, Dept. of IOE, 1205 Beal Ave., Ann Arbor, MI 48109-2117
Chair E-mail:
Chair:
Chair E-mail:

MD43.1 An Investigation into Best Practice Usage of Quality Function Deployment with Proposed Extensions: A US/Japan Comparative Study
• John Cristiano; University of Michigan;

No abstract supplied.

MD43.2 Economic Issues Concerning the Mobility of Scientific Inventions & Implications for Firm Strategy
• Ajay Agrawal; University of British Columbia;

No abstract supplied.

MD43.3 Creative Destruction or Creative Cooperation? Empirical Investigation of Technological Discontinuities & Effect on the Nature of Competitio
• Frank Rothaermel; University of Washington;

No abstract supplied.

# Software Demonstration V

Session: MD45
Date/Time: Monday 15:00-16:30
Type: Software Demo
Track:
Cluster:
Room:
Chair: David Krahl
Chair Address: Imagine That, Inc., 6830 Via Del Oro, Ste. 230, San Jose, CA 95119
Chair E-mail: davek@imaginethatinc.com
Chair: Irvin J. Lustig
Chair Address: ILOG, 25 Sylvan Way, Short Hills, NJ 07078
Chair E-mail: ilustig@ilog.com

MD45.1 Simulation Modeling with Extend

See why Extend is the acknowledged standard for discrete event and continuous simulation in top universities and corporations worldwide. View a demonstration of powerful Extend features: top-down/bottom-up hierarchy, built-in activity-based costing, customizable animation, 1-click output analysis, automatic confidence intervals, sensitivity analysis and integrated development and authoring environments...

MD45.2 Introducing the ILOG Concert Technology

ILOG has recently released new versions of all its optimization components, based on ILOG Concert Technology. This new foundation for the ILOG Optimization Suite provides C++ modeling objects for constraint programming and mathematical programming applications. Come see how ILOG Concert Technology reduces your development time for applications based on ILOG CPLEX and ILOG Solver.

# Data Envelopment Analysis I

Session: MD46
Date/Time: Monday 15:00-16:30
Type: Invited
Track:
Cluster: DEA
Room:
Chair: Yao Chen
Chair Address: Merrimack College, Dept. of Mgmt., Sch. of Bus. & Intl. Commerce, , MA 01845
Chair E-mail: ychen@merrimack.edu
Chair:
Chair E-mail:

MD46.1 Treating Undesirable Measures in Efficiency Evaluation
• Joe Zhu; Worcester Polytechnic Institute, Dept. of Mgmt., 100 Institute Rd., Worcester, MA 01609; jzhu@wpi.edu
• Lawrence M. Seiford; University of Michigan, Dept. of IOE, Ann Arbor, MI 48109-2117; seiford@umich.edu

Undesirable input and output measures are very likely to present. We show that if these undesirable measures are treated as normal measures, the efficiency classification is not valid. Methods on treating undesirable measure are reviewed and a new approach is developed to preserve the convexity in production functions.

MD46.2 An Evaluation of Service Productivity for an Airline Company in Korea using Hierarchical Data Envelopment Analysis
• Young Bum Lee; Ohio State University, Sch. of Public Policy & Mgmt., 314 Fisher Hall, 2100 Neil Ave, Columbus, OH 43210; lee.1491@osu.edu
• Hosun Rhim; Hanyang University, College of Bus. & Economics, Seoul, 133-791 , Korea; hrhim@netial.com
• Kwangtae Park; Korea University, Dept. of Mgmt., , , Korea; ktpark@korea.ac.kr

Service industries such as hospitals and airline companies often have hierarchical organizational structures embedded in them. We discuss the issue of hierarchies of DMUs in DEA and suggest a method that incorporates one level of the hierarchies into the performance evaluation of other levels. We illustrate theHDEA model using data on an airline company in Korea.

MD46.3 Combining Evolutionary Algorithms & DEA
• James Lill; Portland State University, Eng. & Tech. Mgmt. Dept., Portland, OR 97207-0751;
• Timothy R. Anderson; Portland State University, Eng. & Tech. Mgmt. Dept., Portland, OR 97207-0751; tima@emp.pdx.edu

A description of several ways in which information from DEA can be incorporated into evolutionary or genetic algorithms is described and demonstrated using a multi-object knapsack problem. Each individual is allowed to define its own 'efficiency' with respect to the other individuals without the imposition of specific 'niche' parameters.

MD46.4 Productivity & Strategy Shift in Computer & Automobile Industries
• Yao Chen; Merrimack College, Dept. of Mgmt., Sch. of Bus. & Intl. Commerce, , MA 01845; ychen@merrimack.edu
• Agha Iqbal Ali; University of Massachusetts, Isenberg Sch. of Mgmt., Amherst, MA 01003; aiali@som.umass.edu

It is important to reveal strategy shifts at the company level at a particular time and to identify whether the strategic shift is favorable. We demonstrate how the Malmquist productivity index can be decomposed into different components to measure technology and strategy shifts. An empirical study on Global Fortune 500's computer and automobile industries from 1991-1997 is presented.

# Plenary: Improving Performance & Flexibility at Jeppesen: The World's Leading Aviation Information Company

Session: ME44
Date/Time: Monday 16:45-17:45
Type: Plenary
Track:
Cluster:
Room:
Chair: Russell P. Labe
Chair Address: Merrill Lynch, MS Group, Section 31, PO Box 9065, Princeton, NJ 08543-9065
Chair E-mail: russ_labe@ml.com,, rlabe@na2.us.ml.com
Chair:
Chair E-mail:

ME44.1 Edelman Reprise Plenary: Improving Performance & Flexibility at Jeppesen - The World's Leading Aviation Information Company

We introduce a suite of optimization-based decision support tools for production planning and a novel and general method for evaluating and justifying investments in production technology. Our work improved production planning, reduced product lateness and led to improvements in the production processes at Jeppesen Sanderson, Inc., the world's leader in the aviation information industry.

# Decision Analysis at Schlumberger

Session: TA01
Date/Time: Tuesday 08:15-09:45
Track:
Cluster:
Room:
Chair: Gary A. Lundeen
Chair Address: Schlumberger, 8311 North FM 620, Austin, TX 78726
Chair E-mail: lundeen@austin.apc.slb.com
Chair:
Chair E-mail:

TA01.1 Supply Chain Management at a Tester Manufacturing Facility

We developed a LP-based tool for managing the 4-stage supply chain at a tester manufacturing facility. In the presence of long lead times, capacity restrictions and yield losses, this tool was useful for fast computation of ship dates. In addition, it enabled determining effective reactive (to meet quoted shipping dates) measures to uncertainties in the supply chain.

TA01.2 A Study in Risk Attitudes in Schlumberger Geco-Prakla Marine

Geco-Prakla Marine, like all companies, is regularly confronted with the issue of allocating scarce capital and resources among a set of available investment opportunities - opportunities generally characterized by some degree of financial risk and uncertainty. This study was conducted to determine decision maker risk attitudes in order to assist Geco-Prakla Marine management to develop their corporate risk policy.

TA01.3 Life Service Valuation as a Real Options Derivative of Marginal Oil-Field Value

Secondary recovery in maturing oil-fields can be aided by lift services. Many current onshore US fields are either at this stage of their life or will be in coming years. These fields have been an area for electrical submersible pump (ESP) lift services. While an increasing oil-price may make these fields more attractive for ESP services, possible future price downturns carry considerable revenue risk for the service provider...

TA01.4 Simulation of Back Deck Operations on a Marine Seismic Vessel

Marine seismic vessels deploy exploration equipment from their back deck in configurations called streamers. Streamer deployment in a new area, called mobilization, is a complex process and typically takes 4-5 days. This time is considered part of the operational overhead, and is not billed to the clients. Since revenues are measured in tens of thousands of dollars per day, reducing deployment times directly affects profit...

# Tutorial: Multi-Agent Systems

Session: TA02
Date/Time: Tuesday 08:15-09:45
Type: Invited
Track:
Cluster:
Room:
Chair: Abhijit Deshmukh
Chair Address: University of Massachusetts, Dept. of MIE, ELAB 220, Amherst, MA 01003
Chair E-mail: deshmukh@ecs.umass.edu,, http://farm.ecs.umass.edu/~deshmukh
Chair:
Chair E-mail:

TA02.1 Tutorial: Multi-Agent Systems

Multi-agent systems, where loosely coupled decision makers or problem solvers coordinate their activities to tackle a larger problem beyond their individual capabilities, has emerged as a powerful paradigm for representing and solving complex problems. The growth of this field has not only been spurred by the advances in distributed computing and wide spread information connectivity but also by the changing business environment...

# Supply Chain Management

Session: TA03
Date/Time: Tuesday 08:15-09:45
Track:
Cluster:
Room:
Chair: Rachel Q. Zhang
Chair Address: University of Michigan, IOE Dept., 1205 Beal Ave., Ann Arbor, MI 48109-2117
Chair E-mail: rzhang@umich.edu
Chair: Katia C. Frank
Chair Address: University of Michigan, IOE Dept., 1205 Beal Ave., Ann Arbor, MI 48109-2117
Chair E-mail: kfrank@umich.edu

TA03.1 Inventory Policies for Multi-Product Systems with Different Life Cycles
• Katia C. Frank; University of Michigan, IOE Dept., 1205 Beal Ave., Ann Arbor, MI 48109-2117; kfrank@umich.edu
• Rachel Q. Zhang; University of Michigan, IOE Dept., 1205 Beal Ave., Ann Arbor, MI 48109-2117; rzhang@umich.edu

We consider a periodic review inventory system of an item with 2 components. While the lifetime of component 1 is long and assumed infinite, component 2 is discarded after one period. Since demand is correlated among parts, ordering decisions should be made simultaneously. We show that the optimal solution can be described as an (s, S1, S2) policy.

TA03.2 Modeling & Analyzing Vendor Managed Inventory Agreements
• Michael Fry; University of Michigan, IOE Dept., 2811 IOE Bldg., 1205 Beal Ave., Ann Arbor, MI 48109; mjfry@umich.edu
• Tava Lennon Olsen; University of Michigan, IOE Dept., 2811 IOE Bldg., 1205 Beal Ave., Ann Arbor, MI 48109; tlennon@umich.edu
• Roman Kapuscinski; University of Michigan, Business Sch., 8203 Bus. Admin. Bldg., Ann Arbor, MI 48109-2117; kapuscin@umich.edu

We present a model of vendor managed inventory agreements, which are currently popular in industry as a way of introducing cooperation in the supply chain. We examine the effectiveness of VMI agreements in different settings and compare them to other, more traditional forms of supply chain arrangements.

TA03.3 A Dynamic Lot-Sizing Problem with Production Capacity & Demand Time Windows
• Wikrom Jaruphongsa; Texas A&M University, Dept. of IE, College Station, TX 77843-3131; wikrom@tamu.edu
• Sila Cetinkaya; Texas A&M University, Dept. of IE, 238 Zachry Bldg., College Station, TX 77843-3131; sila@ie.tamu.edu
• Chung-Yee Lee; Texas A&M University, 237 Zachry Engineering Ctr., Dept. of IE, College Station, TX 77843-3131; cylee@acs.tamu.edu

We consider a dynamic lot-sizing problem where the production levels are subject to capacity constraints and the customer offers a grace period, a demand time window, during which a particular demand can be satisfied with no early- or late-shipping penalty. We study the optimality properties and provide a polynomial time algorithm for the solution.

TA03.4 No Title Supplied
• Viswanath Cvsa; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106; vxc10@po.cwru.edu

We consider a 2-tier supply chain and analyze the entry decision of a downstream retailer whose costs are influenced by a traded security. We study the investment decision of the potential retailer under 3 pricing policies of the manufacturer including a policy that caps the price of the manufacturer.

# CANCELLED: Value of Information in Dynamic Environments

Session: TA04
Date/Time: Tuesday 08:15-09:45
Track:
Cluster:
Room:
Chair: Joseph Milner
Chair Address: Washington University, 1 Brookings Dr., CB 1133, St. Louis, MO 63130
Chair E-mail: milner@olin.wustl.edu
Chair:
Chair E-mail:

TA04.1 withdrawn - author request of 10/30
• Michael E. Ketzenberg; University of North Carolina, 1019 Laurel Hill Rd., Chapel Hill, NC 27514; mketz@aol.com

TA04.2 withdrawn - chair request of 10/30
• Colin Kessinger; University of Michigan Business School, 701 Tappan St., Rm. 4209C, Ann Arbor, MI 48109; ckess@umich.edu

TA04.3 withdrawn - chair request of 10/30
• Joseph Milner; Washington University, 1 Brookings Dr., CB 1133, St. Louis, MO 63130; milner@olin.wustl.edu
• Panos Kouvelis; Washington University, Olin Sch. of Business, CB 1133, 1 Brookings Dr., St. Louis, MO 63130; kouvelis@olin.wustl.edu

# Stability of Queueing Networks

Session: TA05
Date/Time: Tuesday 08:15-09:45
Track:
Cluster:
Room:
Chair: John J. Hasenbein
Chair Address: University of Texas, Grad. Program in OR/IE, Dept. of Mech. Eng., Austin, TX 78712-1063
Chair E-mail: jhas@mail.utexas.edu
Chair:
Chair E-mail:

TA05.1 Stability & Optimization of Packet Routing in Communication Networks
• David Gamarnik; IBM, Math. Sci. Dept., TJ Watson Research Ctr., Yorktown Heights, NY 10598; gamarnik@watson.ibm.com

In the packet routing problem, digital communication packets have a specified origin and destination in the communication graph and the scheduler has to choose paths along which to process the packets. We construct an asymptotically optimal offline schedule for the static version of this problem and a stable schedule for the dynamic (online) version of the problem...

TA05.2 Establishing Stability for Multi-Class Queueing Networks with Setups
• Otis B. Jennings; Stanford University, Grad. Sch. of Bus., 518 Memorial Way, Stanford, CA 94305-5015; otisj@isye.gatech.edu

In multi-class networks with setups, one cannot ignore questions of stability. We present a general framework for proving stability, provide a heuristic service meta-policy for guaranteeing stability and prove stability of the heuristic when used in conjunction with last-buffer-first-served (LBFS), FBFS, or a specific round robin policy...

TA05.3 Stability of Reentrant Lines with Batch Servers
• Sunil P. Kumar; Stanford University, Grad. School of Business, 518 Memorial Way, Stanford, CA 94305-5015; kumar_sunil@gsb.stanford.edu
• Hao Zhang; Stanford University, Grad. Sch. of Bus., 518 Memorial Way, Stanford, CA 94305-5015;

We explore stability in open reentrant lines with batch servers. We present simple combinations of fluctuation smoothing policies and minimum batch size rules that guarantee stability of reentrant lines with batch servers. We prove stability by adapting a fluid model method, first proposed by Dai.

TA05.4 Scheduling & Stability of Queues with Wait-Dependent Service Times
• John J. Hasenbein; University of Texas, Grad. Program in OR/IE, Dept. of Mech. Eng., Austin, TX 78712-1063; jhas@mail.utexas.edu
• Valerie Tardif; University of Texas, Grad. Program in OR/IE, Dept. of Mech. Eng., Austin, TX 78712-1063; vtardif@mail.utexas.edu
• Elizabeth Campbell; University of Texas, Grad. Program in OR/IE, Dept. of Mech. Eng., Austin, TX 78712-1063; ecampbell@mail.utexas.edu

We consider single-station queueing systems in which the service time of a job may depend on its time spent in queue. We first present some new results on the optimal scheduling policy for such queues. Next, we consider the stability region for these systems under optimal policies. We examine 2 notions of stability, q-stability (queue-length stability) and w-stability (workload stability)...

# Dynamic Traffic Assignment

Session: TA06
Date/Time: Tuesday 08:15-09:45
Track:
Cluster:
Room:
Chair: Srinivas Peeta
Chair Address: Purdue University, Sch. of Civil Eng., 1284 Civil Eng. Bldg., West Lafayette, IN 47907-1284
Chair E-mail:
Chair:
Chair E-mail:

TA06.1 Insights on Stable On-Line Route Guidance Strategies
• Ta-Hui Yang; Purdue University, Sch. of Civil Eng., West Lafayette, IN 47907-1284;
• Srinivas Peeta; Purdue University, Sch. of Civil Eng., 1284 Civil Eng. Bldg., West Lafayette, IN 47907-1284;

Stable on-line route guidance strategies developed using a dynamical system framework and commonly used assignment principles are analyzed. Some real-time implementation issues and solution effectiveness relating to dynamic traffic phenomena are also investigated.

TA06.2 A Combinatoric Algorithm for Single-Destination User-Optimal Dynamic Traffic Assignment
• S. Travis Waller; Northwestern University, Dept. of Civ. Eng., 2145 Sheridan Rd., Tech. Inst., Evanston, IL 60208; travis@trans.civil.nwu.edu
• Athanasis Ziliaskopoulos; Northwestern University, Dept. of Civil Eng., 2145 Sheridan Rd., Tech. Inst., Evanston, IL 60208; a-z@trans.civil.nwu.edu

An efficient combinatoric network algorithm is presented for the single-destination dynamic traffic assignment problem when user-optimal routes are followed. The algorithm is based on the time-space expansion of a network as represented by the cell transmission theory for traffic flow. Algorithm design, properties and computational complexities will be presented.

TA06.3 Estimation of Dynamic Origin-Destination Flows from Traffic Counts in Congested Networks using Bi-Level Optimization
• Hossein Tavana; University of Texas, Dept. of Civil Eng., ECJ 6.2, Austin, TX 78712; hossein@mail.utexas.edu
• Hani S. Mahmassani; University of Texas, Dept. of Civil Eng., ECJ 6.2, Austin, TX 78712-1076; masmah@mail.utexas.edu

In estimation of OD trips from traffic counts, the effect of congestion is often ignored. In other words, link-flow proportions (the proportion of OD flows traversing a link at any time interval) are considered to be constant. However, in congested networks and in particular when drivers have access to real-time route-guidance information, link-flow proportions might vary significantly as demand changes...

TA06.4 Traffic Assignment Properties under Demand Uncertainty
• Jeffrey Fine; Northwestern University, Dept. of Civ. Eng., 2145 Sheridan Rd., Tech. Inst., Evanston, IL 60208; j-fine@nwu.edu
• S. Travis Waller; Northwestern University, Dept. of Civ. Eng., 2145 Sheridan Rd., Tech. Inst., Evanston, IL 60208; travis@trans.civil.nwu.edu
• Athanasis Ziliaskopoulos; Northwestern University, Dept. of Civil Eng., 2145 Sheridan Rd., Tech. Inst., Evanston, IL 60208; a-z@trans.civil.nwu.edu

Insights into the problem of traffic assignment will be examined when the origin-destination demand values are assumed to be random variables with known probability distribution. The objective is to produce properties that can be used to develop more robust solutions for transportation optimization.

# Initiatives in Management & Information Technology Pedagogy

Session: TA07
Date/Time: Tuesday 08:15-09:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Gene Murkison
Chair Address: Georgia Southern University, Dept. of Mgmt., Bus. School, Statesboro, GA 30460-8152
Chair E-mail: murkison@gasou.edu
Chair:
Chair E-mail:

TA07.1 Fact, Fiction or Prediction: An Integrative Exercise for Information Systems Classes

We describe an interactive lecture that demonstrates rapid, and often unforeseen, changes in technology and their impact on decision strategies. Written excerpts from literature, news and textbooks are presented for students to determine whether the excerpts represent fact, fiction or prediction based on the author's intent.

TA07.2 no show
• Yun Wang; Mercy College, Div. of Math/CIS, 555 Broadway, Dobbs Ferry, NY 10522; ywang@mercynet.edu

TA07.3 Internet Teaching Tools for Industrial Engineering
• Reynaldo Rodriquez; St. Mary's University, One Camino Santa Maria, San Antonio, TX 78228-8534; reyrv@yahoo.com
• Rafael Moras; St. Mary's University, One Camino Santa Maria, San Antonio, TX 78228-8534; moras@stmarytx.edu

We discuss the development of a web page that challenges students to solve problems in the areas of manufacturing and human factors. It features games and quizzes, with graphs, photographs and video clips and provides immediate feedback. It can be used as a teaching aid and as a recruitment tool.

TA07.4 The Semester-Long Case in Strategic Management Instruction
• Gene Murkison; Georgia Southern University, Dept. of Mgmt., Bus. School, Statesboro, GA 30460-8152; murkison@gasou.edu

Surveys indicated that a majority of strategic management professors used the small case approach. Capstone courses in business administration require new and innovative techniques across disciplines. I investigated other approaches and analyzed the effects of a semester-long case in real time using MNCs. Future research is suggested.

# Electronic Commerce & Supply Chain Integration

Session: TA08
Date/Time: Tuesday 08:15-09:45
Type: Invited
Track:
Cluster: E-Business/Enterprise Integration: Intl. Bus. Perspectives
Room:
Chair: Li-Chih Wang
Chair Address: Tunghai University, Dept. of IE, Taichung, 407 , Taiwan
Chair E-mail: wanglc@ie.thu.edu.tw
Chair:
Chair E-mail:

TA08.1 An Evolutionary Approach for Developing a Supply Chain Integration System

For any enterprise which wants to develop its SCM system, a systematic developing methodology is necessary and critical. Currently, big-bang and phased-in approaches are employed to implement an integrated SCM system. We present an evolutionary approach which dramatically reduce the risk and cost upon the implementation. The issues discussed include the concept of evolutionary approach...

TA08.2 Guidance on Building Collaborative Commerce
• Jung-Sing Jwo; Tunghai University, Computer & IS Dept., Taichung, 407 , Taiwan; jwo@mail.thu.edu.tw

Collaborative commerce, going beyond rigid supply-chain models and simple information sharing, is considered as the most advanced form of e-business. However, to build a c-commerce application is not trivial. We will introduce our experience of using patterns to help enterprises get into c-commerce quickly.

TA08.3 XML-Based E-Commerce & Supply Chain Integration
• Tai-Ching Tuan; SAIC, 8301 Greensboro Dr., MS E52, McLean, VA 22032; t.tuan@ieee.org

Internet technology makes e-commerce global in nature. Using Web-enabled, text-based XML as a data interchange enabler necessary for business-to-business, e-commerce has grown rapidly recently. We will discuss the issues of leveraging XML to accomplish the e-commerce for procurement and production supply chains automation.

# Military Applications I

Session: TA09
Date/Time: Tuesday 08:15-09:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Cole Smith
Chair Address: University of Arizona, 8858 East Desert Verbena Pl., Tucson, AZ 85715
Chair E-mail: cole@sie.arizona.edu
Chair:
Chair E-mail:

TA09.1 The Programming of Military Acquisitions

The programming of military acquisitions is a resource allocation problem faced during PPBP. It is essentially a multi-period capital budgeting problem. A conceptual design of a DSS is proposed which utilizes a binary integer program to select projects as well as their starting times.

TA09.2 Optimizing Pre-Positioned Strategic Assets for the US Army

We describe the design and implementation of a large mixed-integer programming model used to assist the US Army examine strategic lift requirements, determine the pre-positioning of strategic assets and prescribe flows of troops and equipment to meet a diverse set of deployment objectives from multi-theater warfare to regional humanitarian aid.

TA09.3 Radar Pulse Interleaving for Multi-Target Tracking
• Cole Smith; University of Arizona, 8858 East Desert Verbena Pl., Tucson, AZ 85715; cole@sie.arizona.edu
• Hanif D. Sherali; Virginia Polytechnic Institute & State University, Dept. of ISE, 0118, Blacksburg, VA 24061-0118; hanifs@vt.edu
• Moustafa Elshafei; King Fahd University of Petroleum & Minerals, Dhahran, 31261 , Saudi Arabia;

In a multi-function radar, the maximum number of targets which can be tracked is an important performance measure. Interleaving algorithms developed to operate radars exploit the dead-times between the transmitted and the received pulses to allocate new tracking tasks that might involve transmitting or receiving pulses. We investigate efficient solution methods for the radar pulse interleaving problem...

# New Directions in Scheduling

Session: TA10
Date/Time: Tuesday 08:15-09:45
Type: Invited
Track:
Cluster: Scheduling
Room:
Chair: Nicholas G. Hall
Chair Address: Ohio State University, 301 Hagerty Hall, 1775 South College Rd., Columbus, OH 43210-1144
Chair E-mail: halln@cob.ohio-state.edu
Chair:
Chair E-mail:

TA10.1 An Implicit Enumeration Algorithm for the Batch Scheduling Problem
• Alessandro Agnetis; Universita di Siena, Dipt. di Ingegneria & Info., via Roma 56, Siena, 53100 , Italy; agnetis@dii.unisi.it,, http://www.dii.unisi.it/~agnetis
• Fabrizio Rossi; Universita degli Studi di L'Aquila, Dipt. Matematica Pure e Appl., via Vetoio, Coppito, 67010 , Italy; rossi@univaq.it
• Stefano Smriglio; Universita degli Studi di L'Aquila, Dipt. Matematica Pure e Appl., via Vetoio, Coppito, 67010 , Italy;

Given a set of parts, each requiring a set of tools, the problem is to select the most profitable set of parts compatible with the tool magazine capacity. For this NP-hard problem, we propose an implicit enumeration algorithm based on a particular branching rule. Extensive computational results are reported.

TA10.2 Coordination of Schedules in an Assembly-Type Supplier-Manufacturer System
• Zhi-Long Chen; University of Pennsylvania, Dept. of Systems Eng., 220 South 33rd St., Philadelphia, PA 19104-6315; zlchen@seas.upenn.edu
• Nicholas G. Hall; Ohio State University, 301 Hagerty Hall, 1775 South College Rd., Columbus, OH 43210-1144; halln@cob.ohio-state.edu

In an assembly-type supplier-manufacturer system, the suppliers' production schedules should be carefully coordinated with the manufacturer's schedule in order to achieve a satisfactory system-wide performance. We provide algorithms and complexity results for scheduling problems in this environment. Several efficiently solvable special cases are identified.

TA10.3 Combining Column Generation & Lagrangean Relaxation: An Application to the Common Due Date Scheduling Problem
• Marjan van den Akker; National Aerospace Laboratory NLR, Informatics Div., PO Box 90502, Amsterdam, 1006 BM , The Netherlands; vdakker@nlr.nl
• Han Hoogeveen; Eindhoven University of Technology, Dept. of Math. & Comp. Sci., PO Box 513, Eindhoven, 5600 MB , The Netherlands; slam@win.tue.nl
• Steef L. van de Velde; Erasmus University, Rotterdam Sch. of Mgmt., PO Box 1738, Rotterdam, 3000 DR , The Netherlands; svelde@fac.fbk.eur.nl

The effectiveness of combining column generation with Lagrangean relaxation is demonstrated for an archetypical machine scheduling problem. Our comprehensive computational study shows that the combined algorithm solves instances with up to 125 jobs to optimality, while previous purely column generation algorithm can solve instances with up to only 60 jobs.

TA10.4 Properties of Complementary Hamiltonian Cycles on Bipartite Graphs: An Application in Workforce Leveling
• George L. Vairaktarakis; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106-7235; gxv5@po.cwru.edu
• Daniel Solow; Case Western Reserve University, Dept. of OR & Op. Mgmt., WSOM, 10900 Euclid Ave., Cleveland, OH 44106; dxs8@po.cwru.edu

Consider a paced assembly line where every job is a vector of workforce requirements with as many elements as the number of stations. We study a workforce leveling objective called range. This problem is related to complementary Hamiltonian cycles in bipartite graphs. An efficient solution algorithm is presented.

# Advances in IP Methods & Applications I

Session: TA11
Date/Time: Tuesday 08:15-09:45
Type: Invited
Track:
Cluster: Integer Programming
Room:
Chair: Fred W. Glover
Chair Address: University of Mississippi, Hearin Ctr. for Enterprise Sci, Sch. of Bus. Admin., University, MS 38677
Chair E-mail: fglover@bus.olemiss.edu
Chair:
Chair E-mail:

TA11.1 Heuristics for Boolean Optimization Problems

A BOOP consists in maximizing a linear function in binary variables subject to a Boolean equation. Special cases include set covering, vertex packing and many other combinatorial optimization problems. We describe a heuristic method based on using the best linear approximation of a Lagrangian type combination of the objective function and of a polynomial expression associated to the Boolean constraint...

TA11.2 Branch & Bound Algorithms for the Transportation Problem with Exclusionary Side Constraints

The transportation problem with exclusionary side constraints is formulated as a 0-1 MIP model. B&B algorithms are developed to solve the problem. Penalties are used to strengthen the lower bounds so as to fathom subproblems, peg variables and guide the selection of separation variables. Computational results are reported.

TA11.3 An Optical-Fiber Routing Problem of Synchronous Optical Networks

We consider an optical-fiber routing problem for multiple clusters with minimum cost arising from the deployment of SONETs. We developed 2 mathematical formulations and a B&C procedure for solving the problem optimally. Also, we developed a tabu search heuristic that provides tight upper bounds

# E-Business: An Enabler for the Supply Chain

Session: TA12
Date/Time: Tuesday 08:15-09:45
Type: Invited
Track:
Cluster: Logistics & Supply Chain Management
Room:
Chair: Herbert Heinzel
Chair E-mail: herbert.heinzel@mch20.sbs.de
Chair:
Chair E-mail:

TA12.1 Supply Networks Challenge Enterprise Models

The complexity of business relationships has by far outgrown business processes and structures of today's supply networks. Customers continue to request and expect greater supply network flexibility consistent with their demands. Partner activities have to be synchronized across the network, their infrastructure to be adapted to the complexity of new businesses and customer requests...

TA12.2 Mapping the Supply Chain Operations Reference Model to RosettaNet
• George Brown; Intel Corp., Strategy & Tech. Section;

The SCOR model, developed by the Supply Chain Council, is a methodology positioned to enable competitive supply networks through inter-enterprise business process alignment and performance management. The session describes how SCOR methodology can add value to RosettaNet, a Linqua franca set to standardize electronic communication between supply chain partners in the

TA12.3 Exploring E-Business Options through Supply Chain Simulation

Modeling and simulating supply chain operations enables companies to explore alternative strategies of doing business. Simulation is particularly useful for exploring the relatively uncharted waters of e-Commerce. By comparing the traditional bricks and mortar model with electronic commerce options, companies can achieve a better understanding of the complex interactions and technology associated with e-business...

# Strategic Operations Planning

Session: TA13
Date/Time: Tuesday 08:15-09:45
Type: Invited
Track:
Cluster: Manufacturing & Logistics
Room:
Chair: Sergio Chayet
Chair Address: University of Chicago, Grad. Sch. of Bus., 1101 East 58th St., Chicago, IL 60637
Chair E-mail:
Chair:
Chair E-mail:

TA13.1 Capital Budgeting for Capacity: Options & Operating Risk
• Phillip J. Lederer; University of Rochester, Simon Grad. Sch. of Bus. Admin, Rochester, NY 14627;
• Tushar Mehta; University of Rochester, Simon Sch. of Bus. Admin., Rochester, NY 14627;

We describe a model of capital budgeting where a firm must decide on technology and capacity size. Technology differs in cost structure-the fixed and variable operating cost and initial investment costs. Demand is uncertain. and we evaluate the NPV by adjusting for options value as well as operating risk.

TA13.2 Pricing in Delay-Sensitive Network Markets
• Philipp Afeche; Northwestern University, Kellogg Grad. Sch. of Mgmt., 2001 Sheridan Rd., Evanston, IL 60208;

We study price and capacity decisions in markets where multiple independent providers control interconnected congestible resources. We investigate the impact of congestion effects on the structure of Nash equilibria. This problem is motivated by the situation in commercial data networks, where providers both compete for and jointly serve customers.

TA13.3 Options & Contracting for Capacity Allocation of Capital Goods
• Stefan Spinler; WHU Koblenz, Otto-Beisheim GSM, Burgplatz 2, Vallendar, 56179 , Germany;
• Arnd H. Huchzermeier; WHU Koblenz, Otto-Beisheim GSM, Burgplatz 2, Vallendar, 56179 , Germany;
• Paul R. Kleindorfer; University of Pennsylvania, The Wharton School, 3620 Locust Walk, Philadelphia, PA 19104-6366;

We consider a class of capacity planning problems characterized as expensive, non-scalable, capital-intensive production and pre-commitments of capital payments by multiple agents requiringthe output of this capacity. We propose an options approach to simultaneously determine capacity and to allocate output of the facility to competing agents.

TA13.4 Strategic Capacity Planning under Price & Lead-Time Competition
• Sergio Chayet; University of Chicago, Grad. Sch. of Bus., 1101 East 58th St., Chicago, IL 60637;
• Wallace J. Hopp; Northwestern University, IE/MS Dept., 2145 Sheridan Rd., Tech C210, Evanston, IL 60208-3119; hopp@nwu.edu

We present a game theoretic model of capacity competition in a duopoly where the market clears on price and delivery lead times. We analyze the Stackelberg equilibrium to the game to study the strength of the first mover advantage and investigate options a follower could use to overcome

# Tutorial: Modern Experimental Design I: Planning & Analysis

Session: TA14
Date/Time: Tuesday 08:15-09:45
Type: Invited
Track:
Cluster:
Room:
Chair: C. F. Jeff Wu
Chair Address: University of Michigan, Dept. of Stats. & IOE, Ann Arbor, MI 48109-1285
Chair E-mail: jeffwu@umich.edu
Chair:
Chair E-mail:

TA14.1 Tutorial: Modern Experimental Design I - Planning & Analysis
• C. F. Jeff Wu; University of Michigan, Dept. of Stats. & IOE, Ann Arbor, MI 48109-1285; jeffwu@umich.edu

This tutorial is based on the book 'Experiments: Planning, Analysis & Parameter Design Optimization' by Wu & Hamada (2000, Wiley). The first part will focus on the choice of experimental plans (2-, 3- or mixed-level) fractional factorials and orthogonal arrays) and modeling and analysis techniques. Real experiments are used for illustration.

# Supply Chain Management I

Session: TA15
Date/Time: Tuesday 08:15-09:45
Type: Contributed
Track:
Cluster:
Room:
Chair: A. D. Amar
Chair Address: Seton Hall University, Stillman Sch. of Business, South Orange, NJ 07079
Chair E-mail: amaramar@shu.edu
Chair:
Chair E-mail:

TA15.1 Supply Chain Performance Improvement
• Ali R. Behnezhad; California State University, Dept. of MS, Northridge, CA 91330-8378; ali.behnezhad@csun.edu

A systematic approach to improve the flow of materials through the logistics network will be presented. Integration issues and the role of information technology in the design and management of supply chains will be discussed.

TA15.2 Combined Simulation-Optimization Modeling of Supply Chains: Methods & Applications
• David R. Heltne; Equilon Enterprises, LLC, Westhollow Technology Ctr., PO Box 1380, Houston, TX 77251-1380; drheltne@equilon.com
• Charles R. Standridge; Grand Valley State University, Padnos Sch. of Eng., Grand Rapids, MI 49504-6495; standric@gvsu.edu
• Jeffrey L. Slotter; Shell Chemical Co., 16602 West Kingscoate, Crosby, TX 77532; jls1146@aol.com

We have developed and successfully applied an approach to building combined simulation-optimization models of supply chains. The simulation model represents the operational details while the optimization model represents decision making. Strategic managers and operations engineers use model results in determining capital equipment requirements, inventory capacity and scheduling strategies.

TA15.3 Optimizing the Supply Chain for Business Collaboration
• A. D. Amar; Seton Hall University, Stillman Sch. of Business, South Orange, NJ 07079; amaramar@shu.edu

We provide e-commerce-based models to reduce operating costs of supply replenishments and to shorten their cycle-times. These models are derived from the supply chain system principles. The work also includes models to manage single and multiple location demand centers and provides some exploration on their extension to multiple supply centers.

TA15.4 withdrawn - author request of 10/11
• Shaojun Wang; Louisiana State University, Dept. of IMSE, Baton Rouge, LA 70803-6409; swang2@univx1.sncc.lsu.edu
• Bhaba Sarker; Louisiana State University, Dept. of IMSE, Baton Rouge, LA 70803-6409;
• A. A. M. Jamal; Southeastern Louisiana University, Dept. of Mgmt., 1104 Rue Cannes, Hammond, LA 70402; ajamal@selu.edu

# PROMETHEE Software & Applications

Session: TA16
Date/Time: Tuesday 08:15-09:45
Type: Invited
Track:
Cluster: MCDM
Room:
Chair: Bertrand M. Mareschal
Chair Address: Universite Libre de Bruxelles, Dept. of Stats & OR, Blvd. du Triomphe, CP 210/01, Brussels, B-1050 , Belgium
Chair E-mail: bmaresc@csi.com
Chair:
Chair E-mail:

TA16.1 DECISION LAB 2000: A New PROMETHEE Software
• Bertrand M. Mareschal; Universite Libre de Bruxelles, Dept. of Stats & OR, Blvd. du Triomphe, CP 210/01, Brussels, B-1050 , Belgium; bmaresc@csi.com
• Alexandre Cvetkovic; Visual Decision, Inc.;

Decision Lab is a Windows program that incorporates the PROMETHEE and GAIA methods as well as multiple enhancements such as the treatment of missing values, multiple scenarios and extensive sensitivity analysis and reporting capabilities.

TA16.2 Using PROMETHEE for Antibiotics Selection in an Intensive Care Unit
• Bertrand M. Mareschal; Universite Libre de Bruxelles, Dept. of Stats & OR, Blvd. du Triomphe, CP 210/01, Brussels, B-1050 , Belgium; bmaresc@csi.com
• Jim Ilunga; St.-Michel Hospital;

PROMETHEE is used to assist physicians in selecting the most appropriate antibiotics for the treatment of patients in a Belgian intensive care unit. Criteria such as cost, efficiency, ecological impact, pharmacodynamic characteristics and tolerance are included in the model. The advantages over the traditional efficiency/cost ratio approach are discussed.

TA16.3 Using PROMETHEE to Evaluate Urban Waste Management Systems
• Bertrand M. Mareschal; Universite Libre de Bruxelles, Dept. of Stats & OR, Blvd. du Triomphe, CP 210/01, Brussels, B-1050 , Belgium; bmaresc@csi.com
• Virginie van de Kerchove; Universite Catholique de Louvain, Place des doyens 1, Louvain-la-Neve, 1348 , Belgium;

PROMETHEE is used to asses the performance level of urban waste management systems in several European cities. Economical, environmental as well as social criteria are included in the model. Managers can compare their system with those of other cities, identify their weaknesses, and take appropriate decisions to improve their position.

# Optimization Techniques VI

Session: TA17
Date/Time: Tuesday 08:15-09:45
Type: Contributed
Track:
Cluster:
Room:
Chair: Khosrow Moshirvaziri
Chair Address: California State University, Long Beach, CA 90840
Chair E-mail: moshir@csulb.edu,, http://www.csulb.edu/~moshir/
Chair:
Chair E-mail:

TA17.1 no show

TA17.2 An Approximation Scheme for Fractional Programming
• Jorge U. Villavicencio; Pontificia Universidad Catolica de Chile, Facultad de Matematicas, Vicuna Mackenna 4860, Santiago, , Chile; jvillavi@puc.cl

We present an approximation scheme to solve a certain class of fractional programming problems. The method we propose is based on a logarithmic potential function and it computes an approximate solution to a given accuracy. We show that the number of iterations the method performs is bounded by a polynomial on the given accuracy and on the number of constraints.

TA17.3 A Combined Heuristic Based on Path Relinking & Genetic Algorithms
• Guoqing Zhang; Chinese University of Hong Kong, Dept. of SE/EM, Shatin NT, Hong Kong, , Hong Kong, China; gqzhang@se.cuhk.edu.hk
• Janny M. Y. Leung; Chinese University of Hong Kong, Dept. of SE/EM, 116 Ho Sin Hang Eng. Bldg., Hong Kong, Shatin NT, , Hong Kong; janny@se.cuhk.edu.hk
• Jue Xue; IBM Supply Chain Services, ERP Integration Services; juexue@us.ibm.com

We propose a new heuristic by combining path relinking and a GA. Both parallel and series connections are investigated. We apply the approach to a multiple-level warehouse layout problem. Extensive experiments are carried out to compare the performance of the new approach.

TA17.4 An Efficient Data Structure for Project Networks

An efficient data structure for the nodes-arcs configuration of project networks is introduced. Under this structure, computational time needed to determine critical activities, critical paths and other essential components of the network is greatly reduced. The technique is currently being tested on large-scale networks of construction projects.

# Novel Market Mechanisms & Applications in E-Commerce

Session: TA18
Date/Time: Tuesday 08:15-09:45
Type: Invited
Track:
Cluster: E-Commerce
Room:
Chair: Jan Stallaert
Chair Address: University of Southern California, IOM Dept., Marshall Sch. of Bus., Los Angeles, CA 90089
Chair E-mail: stallaer@usc.edu
Chair:
Chair E-mail:

TA18.1 Pricing Bundles in a Double-Auction Market & their Implications in Business-to-Business Commerce
• Mu Xia; University of Texas, Austin, TX 78712;

An IP model for combinatorial double auctions is proposed, in which complementarities among components are taken into account in bundle orders. A new approach to developing approximate prices for bundles is presented. We show both theoretically and via an example, that it leads to more efficient markets compared to other price-setting approaches.

TA18.2 A Centralized Market for Exotic Derivative Securities
• Siwei Gu; University of Texas, , TX;

A mathematical model is developed to centralize the trading of different exotic derivatives by backing them with a common hedging portfolio. Each derivative is taken as a limit order in terms of its hedging portfolio. The market surplus is maximized and allocated to bidders. By charging a transaction fee, the market maker can cover the rebalancing cost and induce the order that can lower the total

TA18.3 Internal Supply Chain Coordination through Bundle Markets

We present a method of coordination in a supply chain where the activities are coordinated through internal bundle markets. The market mechanism is a double auction where participants can trade bundles of goods.

# Aligning Project Management with Strategy

Session: TA19
Date/Time: Tuesday 08:15-09:45