Knowledge Acquisition Tools and Knowledge Representation Strategies Development for a Naval Expert System. Lessons Learned

摘要

Research performed on the Command Action Team (CAT) system and the Smart Knowledge Acquisition Tool (SKAT) is described. CAT is a rule-based expert system to perform threat assessment in the naval domain; SKAT is an automated knowledge acquisition tool for CAT. These systems were developed jointly by Naval Ocean Systems Center and Carnegie Mellon University. The research explored an alternative knowledge representation scheme for CAT that reduced the number of working memory elements. A knowledge representation mechanism was developed to investigate the feasibility of incorporating higher-level control on rule-firings based on the idea of an expert's problem-solving method. A system is described for automatically generating OPS83 rules from an external database, and an analysis is performed describing SKAT's ability to generate the CAT system. Keywords: Artificial intelligence, Real time. (KR)