Fuzzy knowledge representation and reasoning using Petri nets

摘要

The expert knowledge captured and stored in an expert system makes it possible for nonexperts to solve a particular problem. This expert knowledge is usually expressed in precise form. Many knowledge representation methods have been developed for exact knowledge type. Fuzzy knowledge, on the other hand, is not as easily represented as nonfuzzy (exact) knowledge because it consists of some linguistic (fuzzy) variables. Petri nets, owing to their representing power, have been used to represent exact knowledge. However, many existing Expert Systems use Mycin-like methods to represent uncertain knowledge that assigns a certainty factor value to each uncertain fact. It is possible to use Petri nets to represent fuzzy production rules of a rule-based system. In this article, we propose a modified fuzzy reasoning algorithm based on Chen, Ke, and Chang (1990), which enhances the reasoning capability of the original algorithm. We also propose two additional algorithms for building reachability sets and adjacent places, tables that are required when applying the fuzzy reasoning algorithm. Furthermore, these algorithms are implemented on the Knowledge Craft1 frame-based expert system to build a fuzzy expert system that aims to help students choose their colleges and academic departments.