基于Smith的自适应模糊网络拥塞控制算法  

Network congestion control algorithm based on smith's adaptive fuzzy

作　　者：史永宏[1] 高世界 

机构地区：[1]中国石油大学(华东)计算机与通信工程学院,山东青岛266580

出　　处：《计算机应用研究》2013年第10期3071-3074,共4页Application Research of Computers

摘　　要：针对网络拥塞控制系统中因网络时滞对主动队列管理算法产生的不利影响,提出了一种基于Smith预估的自适应模糊主动队列管理算法。该算法将Smith预估控制与自适应模糊控制相结合,利用Smith预估器补偿网络时滞,同时运用模糊控制在一定程度上克服了传统Smith预估器对模型结构与参数的精确性过于敏感、鲁棒性差等缺点。仿真结果表明,该方法可以使队列长度快速收敛到设定值,同时维持较小的队列振荡,尤其是在网络条件变化的情况下,该算法优于传统PI控制、模糊控制和传统的滑模控制。In order to solve the negative influence of the network-induced delay on AQM algorithm in the network congestion control systems, this paper provided an adaptive fuzzy active queue management （AQM） algorithm based on Smith＇ s predic- tion. The algorithm combined the Smith predictive control and the adaptive fuzzy control Smith predictor compensation the net- work delays ,while using fuzzy control to overcome the less robustness and the excessive sensitive on the accuracy of the model structure and parameters in the traditional Smith predictor to some extent. The simulation results show that the method can make the queue length to quickly converge to the set value, while maintaining a smaller queue oscillation, especially in the case of the network conditions change ,this algorithm is superior to the conventional PI control ,fuzzy control and sliding mode control.

关 键 词：SMITH预估器 拥塞控制 主动队列管理 自适应模糊控制 鲁棒性 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TP393[自动化与计算机技术—控制科学与工程]