Control Theory-based Network Congestion Control

Due to the emergence of many new network applications and rapid expansion of the networked user group, Internet traffic has been increasing dramatically and network congestion appears to be a more and more critical issue. Congestion control mechanisms are very important in guaranteeing the stability of Internet. Moreover, a suitable congestion control mechanism is the basis for other Quality-of-Service(QoS)mechanisms to work effectively.Internet congestion control system can be seen as a nonlinear dynamic feedback system with communication delay. This paper emphasized on designs of active queue management algorithms in network nodes for Internet congestion control. The main research results are as follows:(1) Based on a linearized model of TCP congestion control system, we apply the control theory to convert the design of AQM algorithms to the design of PID controllers using the classically method in frequency domain. Then we implement them in NS.Simulation results verify that the performances of the the PID controller are better than those of PI controller at many aspects.(2) A static output feedback controller is proposed, whose coefficient can be determined by the linear matrix inequality, via giving a robust stabilization theory of the linear state-delayed system. Finally, simulation results indicate a clear advantage of the proposed controller when compared to random early detection and PI algorithms, which also has high robust stabilization under diverse network conditions.(3) A dynamic output feedback controller is proposed, whose coefficient can be determined by the linear matrix inequality based on uncertainty time-delay technology. Simulated results show that the proposed controller is feasible and valid for it possesses not only robustness for all the network parameters but fast settling process.(4) A new proportional-integral-differential controller is established as active queue management AQM scheme to adjust the congestion marked possibility at the links in Internet, whose parameters is tuned by particle swarm optimization algorithm. The simulation results show that the integrated performance of PID controller designed with the new approach is obviously superior to that of the traditional PID controller.