【摘要】 为了评价一种新的主动队列管理(active queue management,简称AQM)算法而改变运行在成千上万个路由器上的软件是不现实的。研究人员被迫在模拟网络或私用网络上测试新算法。分析了网络模拟器NS2的模块组成,并以PSO-PID(Particle Swarm Optimization-Proportional Integral Differential)算法为例,重点描述新AQM算法如何在NS2中实现,详细介绍如何获取AQM算法的评价指标,如平均队列长度、队列抖动、丢包率、链路利用率等定量指标,并结合队列长度变化曲线对AQM算法性能进行评价。更多还原

【Abstract】 It’s impossible to alter the software running on the thousands upon thousands routers to evaluate a novel active queue management(AQM) algorithm.Researchers have to test their algorithms on simulator or private network.The structure of NS2 was first described,and then the implementation of novel AQM algorithm in NS2 was presented with the PSO-PID(Particle Swarm Optimization-Proportional Integral Differential) algorithm as an example.Finally,the AQM algorithms were evaluated in terms of the queue length transformation and the quantity index,such as average queue length,the queue length standard deviation,the packet loss rate,and the link utilization.更多还原