Abstract:

Aiming at the problem that the use of fuzzy Petri nets to make fuzzy reasoning is usually unclear in fuzzy reasoning, a marked fuzzy Petri net(MFPN) is presented to carry out fuzzy reasoning. Several knowledge representation parameters: weight, certainty factor and threshold value, etc, are considered during fuzzy reasoning to obtain more information. A reasoning algorithm based on similarity measures is proposed. The deduced consequence can be calculated according to computing the token value in the final output place. Through the verification of production examples, the meaning of the deduced consequence is clearer and the calculation process is more efficient.

Key words: fuzzy Petri net;fuzzy reasoning;mark;similarity measure;weighted fuzzy production rules