Approximate Reasoning Methods for Decision Aids.
Abstract
This report describes a project of analytic research to develop a decision aid using methods of approximate reasoning. Approximate reasoning, in its most general sense, refers to a set of techniques that attempt to reflect the inherent imprecision in human language and thought. In a stricter sense, it refers to a specific collection of methods subsumed under the theory of fuzzy subsets. In the course of this research effort, a prototype decision aid was developed for Air Force target nomination. Target nomination refers to the process by which Air Force intelligence officers develop an ordered list of targets as candidates for attack. The computer-based decision aid developed combines methods from decision analysis, fuzzy set theory, and computer graphics in order to achieve a system that could facilitate an intelligence analyst's efforts. The basic analytic approach involves two decision-analytic models, an expected value model, and a resource allocation model. The decision aid interacts with the user through a geographical display system. This system provides a map background onto which information about targets may be overlayed. The geographical display system uses videodisc storage and image processing techniques to provide rapid access to a wide range of map coverage.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1982
- Accession Number
- ADA119076
Entities
People
- John F. Patterson
- Jonathan J. Weiss
- Paul J. Sticha