Research And Application Of Chaos Algorithm In Feed-forward Neural Network

In order to understand of information processes in the brain, it is needed to use artificial method to simulate some functions of brain. In previously decades, due to development of neurophysiology, methods have been developed to realize the information processes in brain, such as artificial neural networks (ANN), chaos and so on. All these methods are new edge subjects, so researchers apt to investigate the relations between them.Unlike the gradient descent neural network, the chaotic neural networks has more complex dynamics property, and diversified attractor exist. It is just the dynamics that make it possible for the network to be a technology with abrod application foreground for information processing and optimization computation. A in-depth research is done to chaotic neural network in this paper.Considering the discord of definition of chaos synchronization and different problem in practice, this paper the definition of chaos synchrohization is studied.In this paper , the structure of BP neural network is put forward. The BP algorithm and its steps is also presented.The cognitive ability of multi-layer feed-forword neural network is studied, and some methods are give that can improve convergent speed. Because of the lack of BP algorithm, a chaos learning algorithm are proposed, it is to give one dimension search using chaos and simulated annealing algorithm. It can reach optimization spot.Taking the auto-controlling system of EAF as the object of study, the PID controlling method based on chaotic BP neural network is put forward, and the system design and software development is also presented. Possessing the arbitrary nonlinear expressiveness, the neural network can help to realize the PID control with the best combinations by means of the understanding of the system performance. The network has high precision and good robustness. Results of the EAF control are satisfactory and proves the method to be feasible and effective.