应用多智能体链路认知的低轨卫星网络路由算法

Routing Algorithm for LEO Satellite Network Based on Multi-agent Link Cognition

针对低轨卫星链路不稳定和负载不平衡等因素给网络路由带来的严重影响,提出了一种应用多智能体链路认知的低轨卫星网络路由算法。卫星网络通过多智能体对卫星链路投递率和链路可用性等环境进行感知和推理,获得卫星网络中星际链路质量评价,评价结果用于路由的优化,可达到提高网络吞吐率和使负载均衡的目的。以类似"铱"的卫星系统为仿真对象,对比文章算法和传统的自适应最短路径路由算法在吞吐量、丢包率和端到端时延方面的性能。仿真结果表明:文章提出的算法较自适应最短路径路由算法能增大吞吐量,降低丢包率,缩短高负载时的端到端时延,可有效提高低轨卫星通信网络的路由性能。

Aiming at the serious effects of LEO satellite unstable links and unbalanced loads on network routing,the paper proposes a routing algorithm for LEO satellite network based on multi-agent link cognition.By apperceiving and reasoning satellite link delivery rate and link availability environment by using multi-agent,the satellite network can obtain the interstellar link quality evaluation.The evaluation results are used for routing optimization to improve the network thruput and load balancing.Using a system like Iridium as the simulation object,the algorithm proposed and traditional adaptive shortest path routing algorithm are contrasted in thruput,packet loss rate and the performance of end-to-end delay.Simulation results show that this algorithm proposed can increase the thruput,reduce the packet loss rate and the end to end delay in high load situation,and effectively improve the performance of network routing.