Design Of Optimization And Predictive Control Algorithm For Active Queue Management

At present, both number of users and service type increase rapidly in Internet. The existing network management and control mechanism has not adapted to actual demand, especially when establishing the network conditions of vast scale computing and data processing to satisfy stable and high-efficiency scientific research. So the logical and efficient congestion control mechanism becomes the chief task of network management. Although the service model such as integrated service (IntServ) and differentiated service (DiffServ) and service management mechanism such as admission control, traffic shaping and queue management can satisfy the users' request of quality of service (QoS) at a certain extent, congestion control is still the most essential method. TCP rate control algorithms based on window and its variant algorithms are widely used in Internet as a standard. The complex application's need to QoS is difficult to meet, completely depending on policies and algorithms in the terminal. So it is much more important to study control mechanism of the intermediate node such as router, combining with the terminal algorithms, so as to know the congestion state of the network in time, and actualize efficient resource management policy, accordingly network can avoid congestion efficiently, or come back from severe congestion state as soon as possible.As an enhancement mechanism for the end-to-end congestion control, active queue management (AQM) such as RED can control queue length efficiently on the basis of keeping higher throughput. But research results show that most AQM schemes are very sensitive to the network states such as the variety of link number, service type and time delay. The incorrect parameter setting can often bring queue oscillation, throughput degradation and delay jitter acuity and so on.This thesis presents the background of congestion control and active queue management, and analyses the existed AQM schemes at first. Then the modeling method of TCP+AQM congestion control mechanism based on control theory is described. Next several AQM schemes are proposed based on either numerical optimization or predictive control. Finally the performances of different algorithms are verified and compared in the OPNET platform through the discrete event system simulations. The simulation results show that the algorithms improve the transient performance, and increase robustness against uncertainty such as time delay and variety of number of users, and improve network's utilization.