Accession Number : ADP003923

Title :   An Expert System for Representing Procedural Knowledge,

Corporate Author : SRI INTERNATIONAL MENLO PARK CA ARTIFICIAL INTELLIGENCE CENTER

Personal Author(s) : Georgeff,M.

Report Date : JUN 1984

Pagination or Media Count : 18

Abstract : Development of effective automatic systems for fault isolation and diagnosis requires reasoning about sequences of tests and actions. Unfortunately, traditional expert systems are not well suited to representing such procedural knowledge. In this paper, a scheme is presented that allows the explicit representation of both declarative and procedural knowledge with a unified framework, yet retains all the desirable properties of expert systems such as flexibility, explanatory capability and extendibility. In particular, the scheme allows any heuristic declarative knowledge that maintenance engineers may possess to be integrated easily and uniformly with the strong procedural methods of maintenance plans. Domain-specific metalevel knowledge can also be represented within the same formalism. A simple version of the scheme has been fully implemented and applied to the domain of automobile-engine fault diagnosis. (Author)

Descriptors :   *Systems engineering, *Diagnosis(General), *Automotive vehicles, *Engines, Maintenance management, Heuristic methods, Test equipment

Distribution Statement : APPROVED FOR PUBLIC RELEASE