Study On Chaotic Behavior Of Neural Networks And Their Coupled Phenomena

It is said that human brain is the most complex system in the nature, and human is eager to know its working progress. As the reduction and simulation of biological neural system, the artificial neural network models can be applied to simulate many basic capabilities, which have aroused many researchers'interests. Chaotic dynamics probably plays an important role when brain processes the advanced signals. In this paper, a study is carried out by combining the neural networks and chaotic dynamics.Firstly, this paper describes in detail the origin and development of chaos dynamics, elucidates the significance of the investigation on the artificial neural network. Subsequently, it introduces the theory of artificial neural network and chaotic dynamics, including the concept of chaos and the characteristic of chaos and so on, among them, the Lyapunov stability theories and the Poincaré-map theory are introduced in detail. Thirdly, this article describes that the continuous-time Hopfield neural network constituted in three neuron have chaotic attractor. Finally, we introduce one kind of chaotic Hopfield neural network models and study on their coupled phenomena using Matlab software. It is demonstrated that when several chaotic Hopfield neural networks are connected (coupled) by simple sigmoid signals, the periodic behavior emerges in ensemble of the coupled chaotic Hopfield neural networks. Furthermore, the emergent synchronization among the coupled chaotic Hopfield neural networks is discussed. These are used for studying the progress of information delivery between neural cells.