不确定TCP流模型的离散H_∞鲁棒主动队列管理算法

Discrete-time H-infinity robust active queue management scheme for uncertain TCP flow model

针对TCP/IP网络存在参数时变和不确定性下的拥塞控制问题,提出一种新的基于H∞状态反馈控制的离散鲁棒主动列队管理算法(AQM).该方法针对不确定TCP流模型,将短期突发流所占据的带宽作为系统的外部干扰,同时考虑时滞和参数不确定性因素,基于Lyapunov稳定性理论和线性矩阵不等式技术,设计了离散鲁棒状态反馈控制器以保证路由器队列响应的稳定性和鲁棒性.最后,通过NS-2仿真验证了本文方法的有效性.

Considering the problem of congestion control for the time-varying and uncertain TCP/IP network,we proposed a novel discrete-time robust active queue management（AQM） scheme based on H-infinity feedback control for the TCP flow model with link capacity disturbance and parameter uncertainties simultaneously.In this method,the bandwidth occupied by short-lived connections is treated as the external disturbances,and the effect of both delay and parameter uncertainties is taken into account for the TCP/AQM system model.By using Lyapunov stability theory and LMI techniques,we propose a discrete-time robust H-infinity AQM controller to guarantee the asymptotic stability and robustness of the queue length response of a router queues.Finally the NS-2 simulation results demonstrate effectiveness of the proposed method.