Knowledge representation is a central issue in Artifical Intelligence (AI) research. In order to solve the diverse and complex problems encountered, one needs both a large amount of knowledge and some mechanism for the management and skillful utilization of that knowledge. The basic problem in knowledge representation is the development of an adequate formalism to represent that knowledge. In this thesis I will discuss four of the major techniques for representing knowledge in expert systems: first order logic, production rules, semantic networks, and frames. Using Prolog as the implementation language, I will demonstrate that all of the above mentioned representation techniques, when used in actual implementations, will be reduced to an equivalency - that being a set of Prolog facts and rules. Prolog limits us to a set of facts expressed as "predicate(argumentl, argument, ..., argumentn)" and "IF ... THEN" rules, thus eliminating many of the unique features which characterize the various representation techniques. Therefore, Prolog can be viewed as a representation technique itself.

Library of Congress Subject Headings

Expert systems (Computer science)--Design; Artificial intelligence--Data processing; Prolog (Computer program language)

Publication Date


Document Type


Department, Program, or Center

Computer Science (GCCIS)


Not Listed


Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: QA76.76.E95 H827 1987


RIT – Main Campus