Fang-Yie Leu and Chien-Chiao Yang
Department of Electronic Engineering
National Taiwan Institute of Technology
Taipei, Taiwan, (10772) R.O.C.
In this paper, we propose an effective rule base inspection method that invokes automata theories, graph theories, and AI techniques for rule base consistency and completeness checking by using an algorithmic approach. A knowledge base in this method is represented by one finite state diagram. Through this diagram some errors and anomalies of the rule base can be clearly illustrated and identified. Therefore, rule base maintenance can be easily performed. The characteristics of those states representing such errors and anomalies in the state diagram are investigated. The algorithms to effctively identify those defects are also constructed. Finally, we suggest a sequence for invoking these algorithms.
Keywords: rule base inspection, knowledge-based system, inconsistency and incompleteness
Received April 20, 1990; revised 3, 1990.
Communicated by C. C. Chang.