Chaos Control Method Based On Neural Network Research

Since the early 1990s, the problems of chaos control attract the attention of the researchers and engineers expressly after OGY method found. In this domain, researchers mainly have done two parts of work: theory exploring and engineering applications. There have been some methods of control chaos, such as open-loop control, linear and nonlinear control, intelligent control, OGY method, time-delayed feedback control, fuzzy systems control, and so on.For the excellent performance of nonlinear approximation of neural network, it is fit for controlling nonlinear chaos. The main content of this paper is that control chaos combined with OGY-method or nonlinearity compensation method by nonlinear approximation of neural network. They are incarnated in following words.First of all, this paper sums up the chaos control methods, and catches on the essence of the chaos control, then finds out the ways to control chaos by use neural networks. It also summarizes the method of parameter disturb control and nonlinear compensate control by neural networks, and constitutes the models by mathematics.Secondly, this paper studies the nonlinear approximation by neural network of BPNN, RBFNN, and GA trained. In the OGY method, the neural network approaches the control model. And in the method of nonlinear compensate control, the neural network approaches the nonlinear part of chaos, and use the linear control method to control chaos. This paper controls discrete and sequential chaos respectively by the two methods.Thirdly, based on the simulation by numerical value, this paper analyzes the capability of these methods by convergence, anti-jamming and localization.At last, advances a novel chaos control method by support vector machines algorithm training neural network, and do simulation respectively to discrete and series chaos. The result shows the availability of the new method. And the performance of convergence and generalization show the advantage of the new method.