Fuzzy Logic Based Adaptive Congestion Control Scheme for High-speed Network

This paper proposes a fuzzy logic based adaptive congestion control method to prevent congestion in high speed network such as Internet. The buffer at the router is modeled as a nonlinear discrete dynamic system, and a fuzzy logic system based predictive controller is designed to predict the explicit values of the transmission rates of the sources so as to avoid congestions. Parameter vector tuning methods are provided for the fuzzy logic system weights to estimate the unpredictable and statistically fluctuating network traffic. A Lyapunov-based analysis method is given to demonstrate the stability of the closed loop system. Simulation results show that the proposed method is effective.