Elimination Of Limit Cycles From A Fuzzy Petri Net For Drawing Stable Inferences In Expert Systems

The paper presents a new technique for revision of fuzzy belief of a set of imprecise information embedded in fuzzy Petri net. Limit cycle behaviour, which is often exhibited during the belief revision process in the network, paralyses the mechanism of drawing inference in an expert system,embodying the fuzzy Petri net as the inference engine. The paper provides new algorithms for detection and elimination of limit cycles from a fuzzy Petri net. Instability in the reasoning process in expert systems can therefore be easily eliminated with the help of the proposed technique.