Accession Number : ADP003922

Title :   Application of the CSRL Language to the Design of Expert Diagnosis Systems: The Auto-Mech Experience,

Corporate Author : OHIO STATE UNIV COLUMBUS DEPT OF COMPUTER AND INFORMATION SCIENCE

Personal Author(s) : Tanner,M. C. ; Bylander,T.

Report Date : JUN 1984

Pagination or Media Count : 22

Abstract : Auto-Mech is an expert system which diagnoses automobile fuel systems. Its organization and strategies are patterned after MDX, an expert diagnosis system. The problems that these systems are able to diagnose are represented as nodes within a hierarchy. Each node has knowledge about what nodes to consider next. This approach is intended to be a domain-independent methodology for providing focused problem solving and for localizing knowledge in a conceptually relevant manner. Auto-Mech is implemented in a recently developed language called CSRL, which is specifically intended for building diagnostic expert systems. This paper describes Auto-Mech and discusses why the MDX approach and CSRL were useful in developing Auto-Mech, and where some difficulties were encountered.

Descriptors :   *Systems approach, *Programming languages, *Diagnostic equipment, *Automotive fuels, Fuel systems, Mechanics, Problem solving, Artificial intelligence

Distribution Statement : APPROVED FOR PUBLIC RELEASE