摘要
IPv6定义了一种新的通信模型——选播(Anycast),发送到一个选播地址的报文被传送到由该地址标识的"最近"接口之一。在深入分析选播通信服务模型和遗传算法后,提出了一种多约束的基于改进遗传算法的选播QoS路由算法。该算法采用深度优先搜索和轮盘赌相结合的方法保证初始种群的多样性,引入多种群策略和修正算子,同时对遗传算子进行了改进。仿真实验结果表明,改进后的算法提高了收敛速度,获得了较优的选播路径。
Anycast defined in IPv6 is a new communication model. A packet sent to an anycast that is assigned to more than one in- terface is routing to the "nearest" interface having that address. After anyeast service's model and genetic algorithm are analyzed, improved anycast routing algorithm that has multiple QoS parameters is introduced. The method of Depth_First search and roulette wheel is adopted to make the original population multiple. At the same time, many populations and modification operator are intro- duced and genetic operator is improved. The result of experimental simulation shows that improved algorithm not only can enhance convergence speed, but also can gain better anycast path.
出处
《微计算机信息》
2009年第21期200-201,209,共3页
Control & Automation
关键词
选播
遗传算法
QOS
路由协议
anycast
genetic algorithm
QoS
routing protocol