A sliding mode prediction algorithm for active queue management of TCP networks with time delay

An active queue management (AQM) algorithm based on discrete sliding mode predictive model control is proposed for the network congestion problem presented in internet network system. A sliding mode predictive controller (SMPC) with delay compensation is designed based on the discrete model of dynamic TCP congestion. The current and foretime sliding mode information are used to predict future network dynamics, to correct the sliding mode and to assure the queue convergence to expectations stably and quickly. The asymptotically stability of the time-delay system is proved through using the Lyapunov method. Simulation results in a single bottleneck network system show that the performance of the proposed algorithm is better than the performance of some other algorithms such as PI, RED and SMC in terms of strong stability and robustness. The SMPC algorithm gets over the effects of network uncertainties such as long time-varying time delay, and avoids network congestion effectively.