Chaotic System Control Via Self-constructing Neural Network

Chaos theory is an integral part of nonlinear science, as a natural phenomenon, sometimes, chaos brings disturbance or damage, such as chaos of the power greatly impact on the effectiveness of the system, more over, will cause serious power system crashes. These chaotic phenomena should be controlled and eliminated, Putting forward of OGY control theory raise awareness that chaos can be controlled.Artificial neural network has been proven to approximate arbitrary nonlinear systems, with the characteristics of fitting and generalization, neural networks may control some nonlinear systems; combining the artificial neural networks and chaotic systems, the scholars have made a number of artificial neural network based control method for chaotic systems, in this paper, chaotic systems was controlled with a self-constructed neural network.In this paper, we did deeply study of the neural network control. Control of several types of chaotic system and a backup nonlinear system was achieved well. The main work of this paper:(1) Advanced self-constructing wavelet neural network was put forward. Due to the introduction of the degree measure method (dmm) as well as the selection of hidden layer neurons, the choice of network structure becomes more rationalized and the algorithm becomes more intelligent. We realized the backup system using this controller, simulation results indicate that this controller has a better performance.(2) OGY method,the neural network and the advanced self-constructing wavelet neural network were used to control the chaotic system. Controller design method and stability analysis were given and their control were compared. Then chaotic systems with uncertain was controlled using self constructed RBF network. Simulation results confirm that the controller is reliable.(3) Chaos synchronization is presented based on feedback method as well as synchronization based on neural networks, and the two methods were compared. We give a complete set of control programs, system stability analysis, and simulation verification.