The Gems - Trunk Trouble Analyzer: A Knowledge - Based Expert System for Trunk Maintenance

The Generalized Expert Maintenance System - Trunk Trouble Analyzer (GEMS-TTA) is a knowledge-based expert system prototype designed to diagnose and interpret faults on trunks, indicate repair priorities, and suggest appropriate test or performance monitoring strategies for trunk maintenance. The system interacts with Trunk Operation Center technicians. Experienced technicians use various bodies of domain knowledge to diagnose faults on trunks, to determine real or transient troubles, to test trunks, and to monitor trunk performance. This knowledge is empirical and has been captured in a partitioned knowledge base of the GEMS-TTA. The four segments of the knowledge base are: initial diagnostic knowledge base, initial action knowledge base, extended action knowledge base, and final diagnostic knowledge base. The partitioning is based on the logical interdependence and execution sequence of the bodies of domain knowledge. The expert system uses backward chaining as an inference mechanism to derive conclusions. A prototype of the expert system has been built using KEE,* an expert system-building software tool, on a Symbolics 3670 LISP machine. This prototype represents an effort toward the application of expert systems and artificial intelligence technology in network operation systems for fault diagnosis and facility alarm monitoring.