Research On Active Queue Management Algorithm In TCP Congestion Control

With more and more people being into the Internet, the network data congestion is becoming more and more serious. It is an important problem to solve the network data congestion in the router. In the last few years, active queue management feedback control has become an effective method to solve the congestion of network data.In this thesis, three new active queue management algorithms are proposed, address the data buffer queue problem of network intermediate nodes under dynamic network environment. The global asymptotic stability of the controlled system is analyzed by using the Lyapunov stability theory and La Salle's invariance principle. The effectiveness of the proposed active queue management algorithms are verified by Matlab simulations. The main contents of this thesis are as follows:In the first chapter, the author briefly describes the current situation of the network environment, the network data congestion and the network active queue management control algorithm.In the second chapter, the author gives a brief account of the concept in the analysis of the stability of the nonlinear TCP controlled system, and the mathematical model of the network structure.In the third, fourth, fifth chapters, the proportional-derivative(PD) controller, partial model-based sliding model PD controller, and model-free sliding model-like PD are proposed, respectively. Lyapunov stability theory and the La Salle's invariance principle are employed to show global asymptotic stability. Simulations are presented to verify the effectiveness of the proposed controllers.