Abstract. For the uncertain change characteristics of transition matrix parameters in complex FPN system,this paper presents a parameters optimization algorithm based on CDEDPSO.This article focuses on the weight, threshold and certainty of FPN optimized model and these parameters are analyzed, improved and simulated simultaneously.This method implements dynamic programming and self-learning of uncertain characterization parameters in FPN.Finally, in this paper, the method is applied to a FPN model and the feasibility and effectiveness of this method has been verified. IntroductionThe transition matrix in FPN model is a collection of trigger condition parameters and the biggest drawback is the poor self-learning ability for the uncertain parameters.The changes feature of uncertain parameters are reflected in these areas:the determination of weight, threshold, certainty and other parameters.Therefore,Consider how to deal with and improve uncertain parameters of FPN to optimize FPN uncertain knowledge representation model,enhance the dynamic programming and self-learning ability of FPN and improve the efficiency and accuracy of FPN model has became an important research topic

No abstract

Abstract. In traditional FPN model,the matrix operations are too excessive and it does not meet the requirements of real-time in fault diagnosis system.The paper has improved the FPN model and proposed a CFPN(Controlled Fuzzy Petri Net) modeling method.This method has added controlled place in FPN and improved the design method of control arc of the control place.The method is based on the principle of p-and it enables to simplify the logic of FPN fault diagnosis.That can reduce the matrix computation in FPN fault diagnosis reasoning to improve the efficiency of fault diagnosis.This paper has applied the method in the distillation column fault diagnosis system and verified the applicability and effectiveness of fault diagnosis method based on CFPN model.