The optimal path searching in computer networks using chaotic neural networks with decaying ICMIC

This paper presents a neural network with chaotic dynamics to solve the optimal routing with the reduction of packet loss in computer network. The proposed chaotic neural network (CNN) can control network energy to increase, decrease or keep unchanged through The Iterative Chaotic Map with Infinite Collapses (ICMIC) added to energy function, which can help neural network to enlarge searching space to get optimal solutions and avoid local minima or invalid solutions. The cost function is also defined to represent the cost of optimal path with the reduction of packet loss. In order to verify the effectiveness, the optimal path problem is mapped onto a CNN of two dimensions and then 15-node computer network is optimized for path selection. From the experimental results, the success rate of obtaining optimal solutions of the proposed CNN are higher (3%to 4%) than that of GSNN, and much better (8%to 14%) than that of TCNN.