New Directions in Modeling Tools for Optimization

TD11.1 Bridging the Gap between Spreadsheets & Modeling Languages

• Sam L. Savage; Stanford University, 417 Terman Engineering, Stanford, CA 94305-4026; http://www.stanford.edu/~savage/

Spreadsheets are ubiquitous and interactive, but small, and in the words of Art Geoffrion, 'dimensionally arthritic.' Modeling languages are big and powerful, but understood by very few. Some approaches toward bridging this gap will be discussed.

TD11.2 Remote Access to Optimization Software from Local Modeling Environments

• Todd S. Munson; Argonne National Laboratories, Math & Comp. Sci. Division, 9700 South Cass Ave. Bldg. 221, Argonne, IL 60439; tmunson@mcs.anl.gov
• Elizabeth Dolan; Argonne National Laboratories & Northwestern University, Math & Comp. Sci. Division, Argonne, IL 60439; dolan@mcs.anl.gov

We present the design and implementation of Kestrels, a framework that enables locally installed software packages to access remotely available optimization software via CORBA and pre-existing NEOS Server technology. As examples, we describe Kestrels clients that give existing AMPL and GAMS users easy access to a variety of remote solvers.

TD11.3 New Modeling Language Features for Non-Traditional Optimization Methods

• Robert Fourer; Northwestern University, Dept. of IE/MS, Evanston, IL 60208-3119; 4er@iems.nwu.edu
• David M. Gay; Bell Labs Lucent Technologies, 600 Mountain Ave., Murray Hill, NJ 07974; dmg@research.bell-labs.com

To keep up with developments in solvers for global nonlinear programming and for combinatorial optimization via constraint programming, optimization modeling languages are being forced to become more expressive and to provide a broader variety of information to solvers. We present a range of examples, using the AMPL language for illustration.