Abstract: With the advantages of centralized control and programmable interface, SDN greatly improves the control efficiency and operation flexibility of the network. However, in the deployment and operation of SDN, the high transmission delay, high packet loss rate and other shortcomings are also exposed. To solve these problems, an Adaptive SDN Routing Algorithm based on Genetic Algorithm（ASRAGA） is proposed, which uses genetic algorithm in the global view of SDN to search the optimized path. By constraining the crossover and mutation operation in?the?design, it can avoid the invalid path, reduce the search space scale and the computational overhead of the controller. The ASRAGA adjusts the packet forwarding path by the state of network adaptively. Compared with other algorithms, ASRAGA can improve the network performance and reduce the packet loss rate and network delay on the Mininet simulation platform.

Key words: Software Defined Network（SDN）, transmission delay, packet loss rate, genetic algorithm, routing algorithm

摘要： SDN以集中式的控制、可编程的接口等优点，极大地提高了网络的管控效率及操作的灵活性。但在SDN部署运行中，也暴露出传输时延大、丢包率高等缺点。针对这些问题，提出了一种基于遗传算法的自适应SDN路由算法，该算法利用遗传算法在SDN的全局网络视图中搜索优化路径。算法设计时，对交叉、变异操作进行条件约束，避免产生无效的路径，减小求解空间，降低控制器计算开销。同时能根据网络的动态变化，自适应地选择转发路径。通过Mininet仿真平台进行实验，与其他算法相比，该算法降低了网络的时延以及丢包率，提高了网络性能。

关键词: 软件定义网络（SDN）, 传输时延, 丢包率, 遗传算法, 路由算法