选播是一种网络通信服务,在视频流点播等领域发挥重要作用。该文采用改进的粒子 群算法(PSO)优化QOS 服务,比如端到端延迟,服务器负载,网络流量。该算法采用节点序列编码方案,将离散的组合优化转化为一种准连续的优化问题。实验结果表明,该算法在满足QOS 约束的前提下可以实现选播路由的多目标组合优化。 关键词:选播路由;服务质量;PSO 算法;多目标优化 Abstract: Anycast is a kind of network communication service, and it is vital in domain and so on. PSO algorithm was used to optimization the QoS constraint, such as end – to –end delay, server load and networks traffic. The QoS anycast routing optimization problems were changed into a quasi-continuous problem by using a node series encoding method. The algorithm simulation results show that the algorithm can implement anycast routing combinatorial optimization under QoS constraint. Key words: anycast routing; Quality of service(QoS); PSO; multi-objecttive optimization