An intelligent algorithm for improving QOS in a delay-based end-to-end congestion avoidance scheme

This paper proposes an intelligent algorithm which includes a fuzzy proportional derivative controller to complement a recently proposed delay-based end-to-end congestion avoidance scheme (L. Yan et al., 2004). Aimed to enhance the overall network performance, the proposed fuzzy controller intends to tune control gains based on the stability of the sending rate. The adaptive set point derived assists the controller to stabilize the sending rate so that delay jitter and the loss ratio can be reduced when connections experience propagation delays. The proposed algorithm can work with a wide range of traffic mixtures and propagation delays. The paper demonstrates the improved network performance achieved in comparison with existing algorithms through the use of extensive simulations.