Abstract: In order to utilize network resources with high efficiency and ensure the reliability of the network，the optimum path or routing is always an important part on the research of the dynamic network.But this is a difficult issue no matter whether in theory or in a practical dynamic network.In this paper，a new routing approach based on discrete particle swarm optimization algorithm is briefly discussed to obtain the optimum path between two nodes in the network.Simulation results show that，especially having considered the factors of delay，the proposed approach can be used to search the routing efficiently by the available nodes and paths provided by the reconstructed network topology even if the network changes，and can obtain better optimum convergence property and routing result than ant algorithm.

Key words: routing, reconfiguration, discrete particle swarm optimization

摘要： 寻找最优路由作为动态网络研究的一个重要方面，对于提高网络资源的利用率及可靠性具有现实的应用价值，但无论在理论上还是实际的网络条件下，最优问题一直都是研究难点。针对不同的网络实际条件，提出一种改进的离散粒子群算法来寻找网络中任意两个节点间的最优路由。在以寻找最小路由总延时作为目标函数的情况下，仿真结果显示该算法能较准确地在网络拓扑结构变化的情况下较快地寻找到最优路径，且显示出了比蚁群算法更好的收敛性能，获得了较好的寻优结果。

关键词: 路由, 重组, 离散粒子群