论文部分内容阅读
作为对终端系统上拥塞控制的一种补充 ,中间节点上的主动队列管理 (AQM)策略在保证较高吞吐量的基础上有效地控制队列长度 ,从而实现了控制端到端的时延 ,保证 Qo S的目的 .随机早期检测 (RED)及其派生算法虽可以实现 AQM的技术目标 ,但算法的设计要么是依赖于直觉的 ,要么将网络视为时不变定常系统 ,使得最终形成的算法在稳定性和鲁棒性方面存在不少问题 .在该文的研究中 ,作者将 TCP的流量控制过程等效为二阶线性时变系统 ,采用滑模变结构这种不敏感控制对象参数和负载扰动的控制器设计方法 ,设计出一种新的 AQM算法 .仿真试验表明它的稳态和瞬态性能都优于与它具有相同实现复杂度的 RED和 PI算法 ,并且在负载扰动和参数变化时具有很强的鲁棒性 .
As an addition to the congestion control on the end system, the active queue management (AQM) strategy at the intermediate node effectively controls the queue length on the basis of ensuring higher throughput, thus achieving the control of end-to-end delay and Qo S.The purpose of AQM is to realize random target early detection (RED) and its derivative algorithm, but the design of the algorithm is either based on intuition or treat the network as a constant system, so that the final algorithm Stability and robustness.In the study of this paper, the author equates the TCP flow control process to a second-order linear time-varying system, which adopts sliding mode variable structure which is insensitive to control object parameters and load Disturbance controller design method to design a new AQM algorithm.The simulation results show that both its steady-state and transient performance are superior to its RED and PI algorithms with the same implementation complexity, and in the load disturbance and parameter changes When it has strong robustness.