KNOWLEDGE REPRESENTATION AND KNOWLEDGE MANIPULATION IN DECISION SUPPORT SYSTEMS
Abstract
Decision support systems are information processing systems used in support of the decision making process. The objective is to progress beyond procedural problem processing to a system which generates problem solutions using data bases, algorithms and knowledge systems independent of the particular problem to be solved. A system designer constructs known facts and their interrelationships into a format that allows mechanical processing of the decision maker's problem. The first-order predicate calculus is used as the language of the decision support system. Limitations in the expressiveness of this language are presented and explained. Knowledge is expressed in Horn clause form where the predicates before the implication represent data retrieval and algorithm processing. A format of predicates and terms is presented which facilitates the processing of knowledge by the decision support system. Resolution is used as the problem processing system in this research. The processing proceeds from the problem to be solved through facts and asserted knowledge until a solution is achieved. Differences in resolution for problem solving versus general resolution principles are noted and explained. The major difference is the elimination of complimentary predicates only when all terms in the predicate are instantiated. Strategies for controlling the resolution are discussed and a strategy appropriate for decision support systems is presented.
Degree
Ph.D.
Subject Area
Management
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.