Finding Semantic Errors in the Rule-base of Production Systems, and Reasoning with Insufficient Input Data Petri-net-based Approach

Two simple but practical production systems are modeled using Petri Nets. Petri-net models are very useful in finding semantic errors like generalization error and missing conditions. In addition, reasoning with insufficient input information is formulated to yield meaningful results. In both rule-base verification and reasoning with insufficient data, heuristic methods like backward-chained reasoning and rule modification for transition firing are used. They are, however, versatile enough to be easily adopted in constructing, testing, and operating real and practical production systems. This usefulness is in contrast with other studies which have dealt with mathematical verification or simple application of network and fuzzy theories.