INFORMS Salt Lake City 2000 Sponsor:
Decision Analysis Society

Decision Analysis Arcade: Innovations in Practice

SC01.1 Valuance: Decision Analysis for the Management of an Arbitrary Process

• John Andrews; ;

Valuance is a new framework for applied decision theory, involving more general value-function modeling and a recording system that tracks results in the same terms as projected by the decision analysis model. This provides a generalization of traditional accounting, which remains a special case. We demonstrate with a case study.

SC01.2 Decision & Risk Analysis Applied to the Adaptive Enterprise Concept

• Daniel Owen; Strategic Business Processes;
• Michael Kusnic; Strategic Business Processes;
• Deborah K. Bosch; Strategic Business Processes;

D&RA enables adaptive enterprises by identifying and sensing critical uncertainties and possibilities. Two important D&RA structures follow. 'Potential customer value' (net surplus) has 7 customer-perceivable attributes. Assessments about future customer desires and company abilities become internal free-market signals. Hybrid solutions from rapid, continuous change models are robust under discontinuous change.

SC01.3 Predicting the Decision-Aiding Value of Decision Research

• Rex Brown; George Mason University;

Decision-making generally needs improvement, yet decision-aiding remains primitive. Advances must leverage practitioner experience; however, researchers tend toward what is scientifically attractive rather than useful. Useful research should be rewarded and its practical impact must be cheaply evaluated. A decision theoretic measure is proposed for evaluating, comparing and prioritizing research projects.

Monte Carlo Methods in Decision Analysis

SC02.1 Computational Methods in Decision Analysis: A Comparative Review

• Concha Bielza; Universidad Politecnica de Madrid, Fac. de Informatica, Madrid, 28660 , Spain;
• David Rios Insua; Universidad Politecnica de Madrid, Fac. de Informatica, Madrid, 28660 , Spain;

We review and compare computationally various simulation-based approaches to decision analysis. We consider variants of deterministic methods to account for errors in the expected utility approximation, sample path methods and augmented simulation methods.

SC02.2 Foreign Exchange & Lost Opportunity in the US Department of Defense

• James C. Felli; Naval Postgraduate School, DRMI (64FL), 1522 Cunningham Rd., Monterey, CA 93943-5201;

Using data from the US Air Force and Monte Carlo simulation within a decision analytic framework, we demonstrate how the use of forward foreign exchange contracts and currency options can reduce the financial impact of currency fluctuation for the US DoD.

SC02.3 A Markov Chain Monte Carlo Method for Solving Multi-Stage Decision Problems

• Prakash P. Shenoy; University of Kansas, School of Bus., Summerfield Hall, Lawrence, KS 66045-2003;
• John M. Charnes; University of Kansas, Sch. of Business, 222 Summerfield Hall, Lawrence, KS 66045-2003;

We describe an MCMC method for solving multi-stage decision problems. We start with a valuation network representation that describes a factorization of the joint probability distribution and a factorization of the joint utility function. We decompose the problem into stages and solve each stage using MCMC sampling techniques.

SC02.4 A Specialized Partially Observed Markov Decision Problem Form & Algorithm for Clinical Patient Management

• Niels Peek; Utrecht University, Dept. of Computer Sci., Utrecht, 3508 TB , The Netherlands;

We will describe a special form of POMDP tailored to a particular clinical patient management problem. We will also describe a new solution method based on Monte Carlo simulation for solving such POMDP representations.

Strategy Implementation & Decision Analysis

SD01.1 Including Implementation Uncertainty in Decision Analysis

• Brian Hagen; Navigant Consulting Inc., 2440 Sand Hill Rd., Menlo Park, CA 94025-6900; bhagen@sdg.com
• Mark Brodfuehrer; General Motors, Design Ctr., MC 480-113-A31, 30100 Mound Rd., PO Box 9030, Warren, MI 48090-9030; mark.brodfuehrer@gm.com

Practicing decision analysts often downplay the issues of implementation during the framing and evaluation stages of a decision. Underestimation of what it takes to successfully implement a decision can bias an analysis to the point of making the wrong recommendation. We provide a simple example and discussion of including implementation uncertainty in decision analyses.

SD01.2 Applying Decision Tools when Implementing Strategy in a Union Environment

• Peter J. McAliney; Chenery & Co., 1070 Foster St., Franklin Square, NY 11010; peter_mcaliney@baruch.cuny.edu

Language for decision framing and describing decision tools is technically sophisticated. Conversations within organizations usually involve highly educated senior managers. However, effective implementation with unions often causes leadership and membership to diverge. Decision practitioners therefore must also communicate powerful decision concepts to valued stakeholder groups unfamiliar with the supporting theories.

SD01.3 Putting People into Managing for Value

• Robert W. Lawler; Boeing, 17513 155th Ave. SE, Renton, WA 98058; robert.w.lawler@boeing.com

Improving the way we manage the enterprise, our people and our intellectual capital is essential for achieving full shareholder value. We describe the 3 components and develop a value engine that provides a frame for integrated strategy development and implementation.

SD01.4 The Transition from Decision to Implementation

• Mark Brodfuehrer; General Motors, Design Ctr., MC 480-113-A31, 30100 Mound Rd., PO Box 9030, Warren, MI 48090-9030; mark.brodfuehrer@gm.com
• Brian Hagen; Navigant Consulting Inc., 2440 Sand Hill Rd., Menlo Park, CA 94025-6900; bhagen@sdg.com

One of the most disappointing events for a practicing decision analyst is doing an excellent job in aligning the individuals of a corporation to a specific decision only to find several months later that the implementation of the decision has unraveled. We explore recommendations and experience regarding the successful transition from decision to implementation.

Panel: Resolved - Decision Analysts should be Certified

MB01.1 Panel: Resolved - Decision Analysts Should be Certified

• Ralph L. Keeney; University of Southern California, 101 Lombard St., Ste. 704W, San Francisco, CA 94111; keeneyr@aol.com
• Ward Edwards; Wise Decisions, Inc., 11466 Laurelcrest Rd., Studio City, CA 91604;
• James E. Matheson; Strategic Decisions Group, 2440 Sand Hill Rd., Menlo Park, CA 94025-6900;

Over the years, there have been proposals that some type of accreditation should be available or required for practicing decision analysts. A distinguished panel will discuss the pros and cons of such credentialing. Those attending will be invited to share their views on the subject.

Panel: The Future of Decision Analysis Software

MC01.1 Panel: The Future of Decision Analysis Software

• James S. Dyer; University of Texas, Dept. of MSIS, CBA 5.202, Austin, TX 78712; j.dyer@mail.utexas.edu
• Donald L. Keefer; Arizona State University, Dept. of Mgmt., Tempe, AZ 85287; don.keefer@asu.edu
• Ralph L. Keeney; University of Southern California, 101 Lombard St., Ste. 704W, San Francisco, CA 94111; keeneyr@aol.com
• James E. Smith; Duke University, Fuqua Sch. of Bus., Box 90120, Durham, NC 27708-0120; jes9@mail.duke.edu
• Terry Reilly; Babson College, Math & Sci. Div., Babson Park, MA 02457-0310; reilly@babson.edu

We will address the current state and future potential of decision analysis software. A panel of knowledgeable decision analysts will address such questions as: 'What features or enhancements would add the most value?' and 'What is the appropriate role for software in decision analysis applications?' Audience participation is encouraged.

Decision Analysis & Information Systems

MD01.1 Comparison of Online Auctions & Name your Price to Yield Management

• Sam E. Bodily; University of Virginia, Darden Sch., Box 6550, Charlottesville, VA 22906; bodilys@virginia.edu

Revenue (or yield) management has provided tremendous revenue gains for airlines, hotels, rental car companies and many other companies. Online vendor auctions and 'name your price' may inexpensively capture consumer surplus and potentially do much better than revenue management. Using probabilistic models, we compare the revenue possibilities of these 3 approaches.

MD01.2 Linking Preference with Information Retrieval

• John Butler; Ohio State University, 2100 Neil Ave., Columbus, OH 43210; butlerj@cob.ohio-state.edu

A key assumption for conducting electronic commerce is that the consumer is able to find information that satisfies a set of presumable subjective criteria. We will outline the use of risk value models of preference as the 'scoring' mechanism for evaluating and designing consumer searches of the Internet.

MD01.3 Creating a Decision Support System to Support Production Decisions within the Winter Wheat Supply Chain

• Christopher W. Zobel; Virginia Tech., MS & IT Dept., 2076 Pamplin Hall, Blacksburg, VA 24061; czobel@vt.edu
• Eluned Jones; Virginia Tech., Agriculture & Applied Econ., 321B Hutcheson Hall, Blacksburg, VA 24061; eluned@vt.edu

We discuss the development of a computer-based DSS intended to help wheat producers take advantage of opportunities for greater value-added production. Along with a description of the DSS, an explanation of some of the issues concerned with its development will be given.

MD01.4 ROI? Making Information Technology Decisions

• Robin Dillon; Virginia Tech., 7045 Haycock Rd., Ste. 341, Falls Church, VA 22043; dillon@vt.edu

Popular business literature is constantly complaining about evaluating IT decisions because the 'intangible' benefits of the systems are not captured in an ROI calculation. Is the ROI for information technology decisions an area of application that decision analysis can support?

Symposium: Studying the Ecological Rationality of Simple Heuristics

MD02.1 Simple Heuristics & Military Decision Making

• Kathryn B. Laskey; George Mason University, Fairfax, VA 22030;

No abstract supplied.

MD02.2 Adaptive Decision Making

• Laura Martignon; Max Planck Institute for Human Development, Ctr. for Adaptive Behavior, Lentzeallee 94, Berlin, 14195 , Germany; martignon@mpib-berlin.mpg.de
• Ulrich Hoffrage; Max Planck Institute for Human Development, Ctr. for Adaptive Behavior, Lentzeallee 94, Berlin, 14195 , Germany;
• Daniel Goldstein; Max Planck Institute for Human Development, Ctr. for Adaptive Behavior, Lentzeallee 94, Berlin, 14195 , Germany;

We present analytical and simulation demonstrations of how 7 features of the environment influence the performance of heuristics for choice tasks using binary cues. The features are: inter-cue correlations, center of gravity of cues, number of cues, discrimination rate (combined with validity), compensatory structure, conditional dependencies between cues and training set size.

MD02.3 Simple Heuristics when Driving

• Mike Goodrich; ;

No abstract supplied.

Behavioral Decision Analysis

TA01.1 Measuring Value Trade-Offs

• Philippe Delquie; INSEAD, Blvd. de Constance, Fontainebleau, , France;

We will review how some lessons learned from behavioral decision research can be used to design better methods to measure trade-offs. We will also discuss in what sense these assessment methods are 'better' for decision analysis.

TA01.2 The Surprising Non-Prescriptiveness of Prescriptive Models & How to Exploit It

• Ward Edwards; Wise Decisions, Inc., 11466 Laurelcrest Rd., Studio City, CA 91604;

Behavioral decision research grew from differences between normative and descriptive evaluation models, inference (Bayes) and decision (Max SEU). The 3 normative models together imply decision making is decomposable into 19 subtasks necessary for decisions, with implications for developing decision competence tests and for designing prostheses for the decisionally impaired.

TA01.3 A Behavioral Decision Analysis Approach to Scenario Analysis

• George Wu; University of Chicago, Grad. School of Bus., 1101 East 58th St., Chicago, IL 60637; george.wu@gsbpop.uchicago.edu

Scenario analysis is a widely-used approach for strategic decision making. We explore its psychological usefulness and consider some decision analytic ideas that maximize its value.

Group Decision Making

TC01.1 Virtual Groups & Private Preferences: Social Responsibility in the Information Age

• Kathleen S. Hartzel; Duquesne University, Palumbo Sch. of Bus., 802 Rockwell Hall, Pittsburgh, PA 15282-0180; hartzel@duq.edu
• Nancy Paule Melone; Duquesne University, Palumbo Sch. of Bus. Admin., Pittsburgh, PA 15282; nmelone@nauticom.net
• Timothy W McGuire; Management Science Associates, Inc., 6565 Penn Ave., Pittsburgh, PA 15206-4490; tmcguire@msa.com

We investigate the effects of GSS, meeting structure and anonymity on the individual preferences, group decisions and social responsibility within a social dilemma framework. Personal preferences of majority and minority members were strongly influenced by the face-to-face group process but not significantly influenced by the computer-mediated process.

TC01.2 Guidelines for Mediating Brownfield Cleanup Negotiations

• Jayavel Sounderpandian; University of Wisconsin at Parkside, Dept. of Business, Kenosha, WI 53141-2000; sounderp@uwp.edu
• Nancy Frank; University of Wisconsin, Dept. of Urban Planning, Milwaukee, WI 53201-0413;

Brownfield cleanup projects need cooperation from the current owner of the site, the prospective buyer/developer and the local government. Guidelines are given for a mediator who is charged with finalizing a contingent contract among the 3 parties through negotiation. The guidelines use EU optimization.

TC01.3 Personal Self-Disclosure & Competitive & Cooperative Group Decision Making

• Leah Dietz; Duke University, Fuqua Sch. of Bus., Box 90210, Durham, NC 27708-0210; ledietz@mail.duke.edu
• Susan E. Brodt; Duke University, Fuqua Sch. of Bus., Box 90120, Durham, NC 27708; susan.brodt@duke.edu

We empirically studied how 3-person groups create and distribute resources, after having played a 'getting acquainted' game or not. Results quantified the value of relational concern developed through self-disclosure, revealing an inverse relationship for competitive and cooperative contexts. Self-disclosure heightened self-interest in competitive contexts and heightened communal-interest in cooperative contexts.

TC01.4 Shifts in Willingness to Pay when Individuals or Pairs Face Ambiguous & Unambiguous Risky Choices

• Rakesh K. Sarin; UCLA, Anderson Grad. Sch. of Mgmt., 110 Westwood Plaza, Box 951481, Los Angeles, CA 90095-1481; rakesh.sarin@anderson.ucla.edu
• L. Robin Keller; University of California, Grad. Sch. of Mgmt., 350 GSM, Irvine, CA 92697-3125; lrkeller@uci.edu,, http://www.gsm.uci.edu/~KELLER/
• Jayavel Sounderpandian; University of Wisconsin at Parkside, Dept. of Business, Kenosha, WI 53141-2000; sounderp@uwp.edu

Subjects expressed their willingness to pay for a 50% chance of winning $100 and for a similar ambiguous gamble. They were then randomly paired and each pair expressed its willingness to pay for essentially the same gambles. The responses are analyzed for risky and safety shifts.

Naturalistic Decision Making: An Alternative to Traditional Decision Theory or a Prescription for Disaster?

TD01.1 Panel: Naturalistic Decision Making: An Alternative to Traditional Decision Theory or a Prescription for Disaster?

• Lee Roy Beach; University of Arizona, Coll. of Bus. & Public Admin., Tucson, AZ 85721;
• Robert F. Bordley; General Motors, Renaissance Center, Detroit, MI 48098; robert.bordley@gm.com
• Marvin S. Cohen; Cognitive Technologies, Inc.;
• L. Robin Keller; University of California, Grad. Sch. of Mgmt., 350 GSM, Irvine, CA 92697-3125; lrkeller@uci.edu,, http://www.gsm.uci.edu/~KELLER/
• Gary S. Klein; Klein Associates, Inc.;
• Jim Wise; Eco*Integrations, Inc.;

Naturalistic decision making (NDM) is a new paradigm that studies expert decision making in natural environments. The focus is on rational actors having content expertise making good decisions. The session will introduce NDM along with applications, followed by a panel discussion concerning the pros and cons relative to traditional decision making.

Generalizable Insights from Medical Decision Analysis

TE01.1 Ethics & Medical Decision Making

• Scott B. Cantor; University of Texas, Anderson Cancer Ctr., 1515 Holcombe Blvd., Box 40, Houston, TX 77030-4095; sbcantor@mdanderson.org

Decision analysis and medical decision making both acknowledge the ethical implications of their methodologies. Decision analysis in health care decisions has become a frequent practice since 1975; tradeoffs have become clearer but ethical controversies remain. We discuss contributions of medical decision making to decision analysis, focusing on ethical concerns.

TE01.2 Decision Analysis: Contributions from Medical Decision Making

• Francois Sainfort; University of Wisconsin, Dept. of IE & Preventive Med., 1513 University Ave., Madison, WI 53706; sainfort@engr.wisc.edu

We review how the field of medical decision making has contributed to decision analysis. In particular, we focus on contributions to utility theory development, testing and application. We also discuss current challenges faced by theorists and practitioners of medical decision making.

TE01.3 Stochastic Trees in Medical Decision Modeling

• Gordon B. Hazen; Northwestern University, IEMS Dept., Evanston, IL 60208-3119;

Stochastic trees are a recent modeling innovation for medical decision analysis which graphically combine decision trees and continuous-time Markov chains. Stochastic trees can usually be factored into simpler components, a process which eases model formulation and presentation. Applications to medical cost effectiveness and medical decision analysis will be presented.

TE01.4 Medical Decision Analysis: Lessons Learned

• Arthur S. Elstein; University of Illinois, Dept. of Medical Education, 808 South Wood St., MC 591, Chicago, IL 60612-7309; aelstein@uic.edu

Decision analytic methods, including clinical guidelines and evidence-based medicine, were introduced into clinical medicine in response to concerns about practice variation and cost-effectiveness. Simple models, while more acceptable to physicians, don't satisfy professional standards. Complex models, however, risk disempowering physicians or being ignored. These tradeoffs depend on who wants the analysis.

Generalizable Lessons from Government Decision Analysis

WA01.1 R&D Portfolio Analysis for Department of Energy Environmental Management

• Gregory S. Parnell; US Military Academy, Dept. of Systems Eng., Mahan Hall, Rm. 342, West Point, NY 10996; gparnell@usma.edu
• Sheila Jordan; SAIC, 9 East Second St., Frederick, MD 21701; sjordan@unitec-md.com
• David Geiser; Department of Energy, EM Office of Sci. & Tech., Cloverlead, Germantown, MD; david.geiser@em.doe.gov

The DoE's Office of Science & Technology is responsible for an annual budget of more than $250M. We describe the R&D portfolio analysis we have successfully used for the past 3 years to help environmental management decision-makers select the best R&D portfolio. We also discuss lessons learned.

WA01.2 Use & Non-Use of Decision Analysis in a Major Department of Energy Decision at Hanford

• David A. Seaver; Battelle Pacific Northwest National Laboratory, PO Box 999, MS K8-03, Richland, WA 99352; david.seaver@pnl.gov
• Andy Hesser; Battelle Pacific Northwest National Laboratory, PO Box 999, MS H6-61, Richland, WA 99352; andrew.hesse@pnl.gov
• Mark Robershotte; Battelle Pacific Northwest National Laboratory, PO Box 999, MS H6-61, Richland, WA 99352; mark.robershotte@pnl.gov

In 1998, the DoE made a multi-billion dollar decision to authorize a privatized contract to process nuclear waste at Hanford. The decision, decision-analytic concepts used to structure the decision, some specific analyses and lessons learned regarding the application of decision analysis in a highly visible, political decision are discussed.

WA01.3 Selecting an Airport Vulnerability Assessment Methodology

• Rick Lazarick; Federal Aviation Administration;

We address the use of decision science methods in the evaluation and selection of a methodology for evaluating the vulnerability of airports to terrorist acts. The FAA and its assembled Blue Ribbon Panel rigorously employed the evaluation methods and reached a conclusive decision on the most desirable methodology.

WA01.4 Multi-Attribute Risk Analysis in Nuclear Emergency Management

• Raimo P. Hamalainen; Helsinki University of Technology, Systems Analysis Lab., PO Box 1100, Espoo, 02015 , Finland; raimo@hut.fi
• Mats R. K. Lindstedt; Helsinki University of Technology, Systems Analysis Lab., PO Box 1100, Espoo, 02015 , Finland; mats.lindstedt@hut.fi
• Kari Sinkko; Radiation & Nuclear Safety Authority (STUK); kari.sinkko@stuk.fi

The Finnish radiation protection authorities practiced MAUT-based risk analysis in the early phase decisions on countermeasures after a simulated nuclear accident. The goal was to deal with conflicting objectives, different parties involved and uncertainties inherent in such crisis situations. This study was part of the EU-RODOS project.

WA01.5 Decision Analysis for Evaluating Environmental Regulation: Cautionary Tales

• Rex Brown; George Mason University;

Decision analysis attempts to evaluate the cost-benefit of proposed and past regulations realistically often run afoul of government and business politics. Examples and suggested remedies are drawn from consulting experience, including a Congressional mandate to EPA to determine if the Clear Air Act has been (or will be) worth its cost.

DA 2000 & Beyond: How Industry is Learning & Applying Decision Analysis

WB01.1 Decision Analysis in the 21st Century: Integrated Decision Management

• James McCuish; BP/Amoco, 3700 Bay Area Blvd., Houston, TX 77058;

Integrated Decision Management enhances decision analysis and real options methodologies by actively linking them into the organization's systems and cultures for value engineering, project management and capital portfolio optimization. Ignoring this often creates an 'orphan status' for decision analysis that does not deliver timely results or long-term value.

WB01.2 Decision Analysis in the 21st Century: Integrated Portfolio Management

• Gerald A. Bush; Decision Strategies, Inc., Atlanta, GA 30350; gabush@decisionstrategies.com

The drive for value and R&D efficiency has led to new thinking about the principles and methods for evaluating pharmaceutical projects using probabilistic discounted cash flow and real options methods. Integrated Portfolio Management incorporates the best practices and software automation into a modern system for today's leading companies.

WB01.3 Decision Analysis in the 21st Century: Pharmaceutical Portfolio

• Cynthia S. Sushak; Novartis Pharmaceuticals, 59 Route 10, Bldg. 122/336, East Hanover, NJ 07936; synthia.sushak@pharma.novartis.com

The application of Integrated Portfolio Management to a pharmaceutical portfolio led to a significant R&D savings and enhanced understanding of each project's potential and risk. The approach used probabilistic cash flow and real options methods to appropriately value each type of project. Portfolio Wizard software enabled quick optimization and display of scenarios.

WB01.4 Decision Analysis in the 21st Century: Multimedia Training

• David C. Skinner; Decision Strategies, Inc., 13410 Queensbury Lane, Houston, TX 77079; skinnerdc@aol.com

A series of hands-on workshops combined with Internet-based tutorials are available to learn Integrated Decision Management. The spectrm of concepts and skills ranges from the fundamentals for general audiences to advanced topics for practitioners. Innovative use of multimedia technology adds depth and individual flexibility to the learning process.