The Optimization And Application Of Fuzzy Neural Petri Nets Model

In recent years, Petri nets has been widely used and researched by scholars, as a strong modeling tool for information processing system, which have characteristics of discrete, asynchronous, concurrent, parallel, non-deterministic and stochastic. Moreover, fuzzy theory is much favored by oriental scholars as its advantage of uncertain information representation. As the same time, because of its own strong nonlinear processing capability neural network has been successfully utilized in many kingdoms, such as computer science, neuroscience and economics etc.Fuzzy Neural Petri nets (Fuzzy Neural Petri nets, FNPN) is proposed by combining Petri net modeling capabilities with fuzzy representation, introducing adaptive neural network algorithm. FNPN brings together advantages of three theories respectively, then it can solve problem of some fuzzy parameters which are difficult to determine in the model, model is more truly describe actual system. But its ability depends on learning algorithm selected, if you want to get a good model, you need to use speed learning convergence, high accuracy and small shocks learning algorithm. In summary, to improve the optimization algorithm of FNPN model has a strong practical significance to establish a more ideal model.For poor computational accuracy, convergence and large network concussive during training of fuzzy neural Petri net learning algorithm, an optimization algorithm was proposed. Two S-type continuous functions were used to express transitions enabling and the new tag values after transitions firing respectively, in addition, the value before corrected was considered, then new momentum was added based on the traditional parameters correction formula. It was proved that the optimized parameters correction algorithm can ensure the convergence of the FNPN network. In the application of the improving algorithm, this article will apply it in fault diagnosis system that has been widely researched, through establishing FNPN model of expert systems and fault diagnosis system, using the optimized algorithm to calculate the model, we can see that the effect of the algorithm is more excellent than general algorithm at accuracy, efficiency and the convergence of algorithm from data charts.