Abstract:

It is significant and being unsolved yet for building a Fuzzy Petri Net (FPN) so as to get rid of the shortcomings of poor selflearning ability. To address this problem, differential evolution algorithm is originally introduced into the procedure of exploring parameters of FPN. According to the actual characteristics of FPN, an improved differential evolution algorithm is proposed. The algorithm utilizes the chaotic strategy to generate initial population and integrates selfadaptive factors with precocious punishment strategies as a result of enhancing the diversity of population, while ensuring being strong convergent and global. Simulation experiment shows that the trained parameters gained from the proposed algorithm are 5 times accurate than any other traditional algorithms.   

Key words: fuzzy Petri net(FPN);fuzzy reasoning;improved differential evolution algorithm;precocious punishment