Research On Chaotic Neural Network And Its Application

Artificial neural network is a new information processing system based on the human brain structure and the activities mechanism. Artificial neural networks can behave many characteristics of the brain and have some of the basic functions of human brain, imitating the neural frame of human and the mechanism of many activities. Recently, it is found that there are many chaotic phenomenon of chaos in the nervous system, so the research of chaotic neural network (CNN) is becoming a new task for us.This paper studied the chaotic neural network and applied it to a typical combinatorial optimization problem and broadband matching network design. The main content of this paper can be divided into two parts: first, the theory introduction of chaos, the classical Hopfield neural networks and the chaotic neural network; second, the applications of the chaotic neural network. Specifically, the following sections are included:(1) Introduce completely the basic theories of the chaotic dynamics, giving the concept, basic term and basic characteristics of chaos. Analysis the relationship between the Lyapunov index and chaotic state through a typical chaos map——Logistic map. The basic chaotic algorithm had been proved to be effective by a typical function.(2) Summarize the different kinds of chaotic neural network, modify the transient chaotic neural network (TCNN) model and apply it in simulation of complex nonlinear function. The simulation results proved that the improved transiently chaotic neural network was more effective than the classical Hopfield neural network.(3) Explain the energy function of Travel salesman problem (TSP) by Hopfield neural network, and solve 10-citys TSP with Hopfield neural network and Transient chaotic neural network. The simulation results proved that the transient chaotic neural network was superiority to solve TSP problems.(4) Introduce the development of broadband matching theory and the real-frequency method used in broadband matching. This paper proposed a method to design broadband matching network based on TCNN, give the objective function of a practical antenna and optimized it, and then integrated out of the antenna matching network by the optimization results. The simulation curve showed this matching network was effective.