基于混沌优化的大时滞网络拥塞控制算法  

Congestion Control Algorithm Based on Chaos Optimization in Large-delay Networks

作　　者：侯萍[1,2] 王执铨[1] 

机构地区：[1]南京理工大学自动化学院,南京210094 [2]南京人口管理干部学院,南京210042

出　　处：《系统仿真学报》2009年第17期5486-5490,5497,共6页Journal of System Simulation

基　　金：国家自然科学基金(60574082)

摘　　要：基于改进的网络简化模型,提出一个大时滞网络拥塞控制算法称为改进的混沌优化PID控制算法(ICPID)。改进的网络简化模型考虑了时滞的影响,利用这个模型,AQM路由器应用改进的混沌优化策略优化PID控制器参数。通过加速混沌搜索快速得到控制器参数的次优值,运用变尺度方法在次优值附近找出控制器参数的全局最优值。对比仿真结果表明,该算法有较好的鲁棒性,链路利用率更高,丢包率更小,平均队列长度更趋于期望值,综合性能优于PI和DC-AQM。A new congestion control algorithm called PID controller design with improved chaos optimization algorithm based on improved model with delay term was proposed in large-delay network situations. According to the model, the PID parameters were tuned with chaos optimization in AQM routers. This optimization algorithm that mixed stepped-up chaos optimization algorithm and mutative scale strategy could speed up the chaos ergodic optimum searching of the controller parameters. Compared with PI and DC-AQM via experimental simulations, the proposed method is more efficient and robust in achieving the lower packet loss rate and higher link utilization.

关 键 词：主动队列管理 网络拥塞 PID控制 混沌优化 大时滞 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]