Robust Adaptive Active Queue Management Algorithm On Dynamic Network Systems

In the rapid development of network technology today, the network system has been widely used, the network has gradually expanded the scale of the network system is also gradually increasing congestion. When into line with more than one packet arrival and the need for the same output line, or router processing speed slow, then would lead to network congestion. Congestion would cause network data transmission delay, further increasing congestion, the network will have a serious crash. However, if only relying on the source side of the flow control strategy, then the utilization of network resources will be greatly reduced, while the network congestion problem can not be a good solution. Active Queue Management (AQM) algorithm can be a good solution to network congestion problem, but this method is gradually used in the network system. AQM queue length by maintaining the system in a stable context, raise the router packet queue utilization and reduce system latency and improve network performance, increase network quality. Therefore, the AQM algorithm is of great significance to the study.Control theory is applied to active queue management algorithm in this article, the proposed three kinds of active queue management algorithms, network by controlling the queue length and improve the delay on the sector, improve the stability of the system. The contents and conclusions of the study are as follows:First of all, the existing network system model is improved at the same time network system H_∞controller of the network system is designed to compensate time-delay for improving the upper bound of the network system time-delay. By selecting Lyapunov function to analyze the system stability and system H_∞performance.Second, in the case of the state of network is unmeasurable,the H_∞controller of network system's robust H_∞output feedback is designed. In the controller designed, the application of integral equation theorem and the introduction of a free matrix can be a very good network system which regulate the queue length, reducing outside interference and and a number of uncertainties brought about the harm to the network, so network The queue length remain near the equilibrium point, effective to avoid the occurrence of network congestion, improve the network performance of the systems. Finally, a kind of adaptive controller with memory is designed which can update the parameters in real time based on network congestion, parameters during operation of the network congestion can be self-adaptive change. During the simulation, the adaptive controller parameters can be adjusted well and the queue length of network system can be maintained at equilibrium point.