A Modified Gert Network For Automatic Acquisition Of Temporal Knowledge

L D. Interrante
J E. Biegel

Abstract

This paper describes a technique which has been developed for automatically acquiring temporal knowledge for use in an Intelligent Simulation Training System (ISTS)*. The objective of the ISTS is to train students in monitoring and controlling physical objects in time and space. The ISTS consists of a graphic computer simulation, an expert system, and a user interface. The graphic computer simulation represents the system to be monitored. The trainee responds to routine and/or critical situations as depicted in the simulation by typing in commands to simulation objects. ISTS domain knowledge is automatically acquired by a subsystem known as ELICIT: Expertise Learner and Intelligent Causal Inference Tool. ELICIT incorporates a modified GERT network for representing knowledge concerning dependencies among temporal events. The network is used as a basis for prompting the domain expert for knowledge. In addition, ELICIT draws upon knowledge contained in the network for the development of contingency plans concerning timing of events. This implementation is coded in Common LISP and Joshua (a knowledge representation language developed by Symbolics, Inc.) on a Symbolics 3630 LISP machine.