聚合组播优化中的蚁群算法研究  

Research on Ant Colony Algorithm for Aggregated Multicast Optimization

作　　者：伊善文[1] 王华[1] 于超英[1] 

机构地区：[1]山东大学计算机科学与技术学院,山东济南250101

出　　处：《小型微型计算机系统》2010年第10期2043-2048,共6页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60773101)资助

摘　　要：当大量组播组并存于网络中时,IP组播将遭遇严重的组播状态扩展性问题.聚合组播是针对该问题的一种新颖的解决方案,它的主要思想是使多个组共享同一棵聚合树,从而减少组播转发状态.树选择问题作为聚合组播的核心问题已经被证明是一个NP完全问题.本文提出一种改进的蚁群算法(ACAM算法)对聚合树进行选择.仿真结果表明该算法在聚合度、转发状态降低率等性能指标上都优于传统聚合组播算法.IP multicast faces a serious state scalability problem when there are large numbers of groups in the network. Aggregated Multicast has been proposed as a novel solution to solve this problem, in which multiple groups shared one aggregated tree, so as to reduce the multicast forwarding states. The key idea in Aggregated Multicast is tree selection, which has been proved NP-complete. In this paper, an algorithm ACAM （Ant Colony for Aggregated Multicast） is proposed to select aggregated trees. Simulations have shown that this algorithm has much better performance than tradition algorithms both in aggregation degree and the state reduction ratio.

关 键 词：聚合组播 最小集合覆盖 树选择 蚁群算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]