基于混沌神经网络的最短路径路由算法

Shortest Path Routing Algorithm Based on Chaotic Neural Networks

将最短路径问题映射到混沌神经网络,提出了一种带有混沌噪音的神经网络最短路径路由算法。首先设计了与最短路径有关的网络费用和路径表达方法;其次结合混沌神经网络的数学模型建立神经元的运动方程;最后依据网络费用和约束条件构造神经网络的能量函数。分别在具有9个结点和15个结点的网络拓扑结构上进行了实验,单个和多个分组请求均能快速地找到最短路径。结果表明,本文提出的最短路径路由算法用于高速交换网络是有效可行的。

Representing the shortest path problem as a combinational optimization problem, this paper presents a shortest path routing algorithm based on CNN. The expressions of network cost and shortest path are given. The motion equations of neurons ave obtained by means of the model of CNN. An energy function is constructed which takes into account all objectives and constraints. It applies the CNN algorithm to 9-node and 15-node network model, respectively. The shortest path could be found for single and multiple packets request. Simulation results show it validating that using chaotic noise neural network to solve the shortest path problem in a communication network.