基于人工神经网络的分簇多播路由算法  

Multicast Routing Algorithm Based on Clustering Structure by ANN

作　　者：方向[1] 冯烟利[1] 

机构地区：[1]山东工商学院计算机系,山东烟台264005

出　　处：《微电子学与计算机》2010年第5期78-81,86,共5页Microelectronics & Computer

基　　金：山东省自然科学基金资助项目(Y2006G13)

摘　　要：为了能够快速有效地求解多播路由模型,融合分簇算法和多播树算法思想,利用神经网络作为计算工具,提出了一种基于分簇思想的多播路由算法:用自组织神经网络来分簇,用Hopfield神经网络来构造一个由簇首节点组成的多播树.该算法不但简单易行,而且具有收敛速度快、效率高、控制开销小的特点.最后,通过仿真从实践上支持了本算法的可行性.With the purpose to solve the Multicast routing model efficiently,we combine clustering methods and multicast tree algorithms,a cluster-based multicast routing algorithm is proposed. The problem is tackled by ANN：A self-organising-map neural network has been used for clustering; Hopfield neural networks are used to construct a multicast tree,whose nodes are all of cluster-head. The algorithm is not only simple and characterized with its quick convergence,efficient and low maintenance costs of routing. Meanwhile,the efficiency of the method is supported by Simulation.

关 键 词：人工神经网络 分簇 路由 多播树 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]