With Fuzzy Bp Algorithm, Neural Network In Nonlinear Dynamic System Identification

This paper mainly investigates the application problem of fuzzy network and neural network with BP algorithm in nonlinear dynamic systems identification. By comparing their advantages and disadvantages, we propose the identification methods of these two systems respectively. Finally, the simulation examples verify the effectiveness of BP network in fuzzy network identification and neural network identification. The main results are as follows:The first part is the introduction of this paper which surveys the development of the system identification, the methods of traditional system identification, and discusses two methods which are fuzzy identification method and network identification method respectively. Furthermore, we discuss their main idea, improvement version, and their advantages and disadvantages, and give the theory of BP network for the system identification.In the second part, we suggest a corresponding back-propagation learning algorithm with the help of describing the fuzzy logic systems as feed-forward network with three layers, view the fuzzy logic systems with this BP study algorithm as the identifier of nonlinear dynamic systems.In the third part, in terms of the theory that forward neural network can approach any continuous function defined in the compact aggregation, we view the BP neural network as the identifier of nonlinear dynamic systems, and accurately identify all kinds of nonlinear dynamic systems.