Learning Algorithm with Fuzzy Petri Nets Model Based on Taboo Search

All parameters of fuzzy Petri nets are hard to be determined and learning mechanism is lacking. In this paper, a learning algorithm with fuzzy Petri nets model based on taboo search is proposed to solve the problem. The paper gives two basic kinds of fuzzy Petri nets model for the representation of the fuzzy production rules. And several fuzzy reasoning functions are built for fuzzy Petri nets model. Compared with other fuzzy reasoning algorithm, it is suitable to fuzzy Petri nets learning. Taboo search is introduced, the index of which is mean squared error of terminate place´s real confidence and anticipant confidence. By the cooperating of the neighborhood generation and the strategy of intensification and diversification, parameters learning of fuzzy Petri nets are guided. The simulation example shows that the resultant fuzzy Petri nets model owns strong generalization capability and self-adaptability.