Competitive Tradeoff Modeling: Methodology, Computation and Testing
Abstract
The primary aim of this research program was to investigate ways to help people to make better decisions about tradeoffs in resource allocation in situations frequently involving uncertainty, with particular application to Army problems in force mix and weapon mix issues. One of the main areas was the computational implementation of large-scale scenario analysis models. Such models employ a finite probability distribution describing different scenarios (i.e., possible states of the decision environment), and they then compute solutions that optimize against these scenarios according to some appropriate decision criterion. For example, one might seek a decision that provided maximum expected performance on some measure of effectiveness, subject to various operational constraints (which can be different in different scenarios), or one might look for a decision providing minimum variance in performance across scenarios, subject to a floor on expected performance. The main practical result achieved from this work was computational implementation of the so-called bundle-trust region (BTR) method and its testing on numerous scenario analysis problems, including force mix problems furnished by the U.S. Army TRADOC Analysis Command (TRAC). This report describes the problems studied under the subject contract. Accomplishments are reported in force modeling, performance evaluation, and related supporting areas of optimization. Papers produced under the contract and scientific personnel participating in it are listed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1992
- Accession Number
- ADA260439
Entities
People
- Stephen M. Robinson
Organizations
- University of Wisconsin–Madison