基于神经元自适应变结构控制的主动队列管理算法

针对网络TCP模型的非线性以及回路延时和负载波动等不确定性因素，提出一种基于神经元自适应变结构控制(VSC)的主动队列管理(AQM)算法。通过非线性变结构控制以保证路由器队列响应的快速性和鲁棒性；同时考虑到滑模控制中存在的抖振会引起队列波动和控制精度降低等问题，引入神经元在线调整控制器参数以减弱抖振，从而减小队列延时和模型不确定性的影响，提高AQM系统的鲁棒性和性能。最后通过NS-2仿真实验验证了算法的有效性。

Active queue management algorithm based on neuron adaptive variable structure control

Considering the non-linearity of TCP model, uncertainty of Round Trip Time (RTT) and fluctuation of network load, an Active Queue Management (AQM) scheme based on Variable Structure Controller (VSC) using single neuron adaptive learning was proposed. The nonlinear VSC was used to guarantee the swiftness and robustness of queue response at router. However, the jitter of VSC would cause the queue fluctuation and performance degradation. Therefore, a single neuron was introduced to adjust the parameters of the VSC in order to alleviate the effect of jitter and modeling uncertainty. The proposed scheme can reduce the jitter and enhance the robustness for AQM control system greatly. Finally, the simulation results show the effectiveness of the proposed algorithm through NS-2 simulator.