Self-tuning fuzzy traffic rate control for ATM networks

This paper addresses the problem of rate control for the available bit rate (ABR) service class in an ATM network. A Sugeno's method based fuzzy ABR rate controller is proposed in this center. This fuzzy congestion control approach is optimized by a conventional gradient search method, in which the parameters of the output fuzzy functions are adapted based on the gradient of a certain control criterion. By simulation using real trace data we show that ATM networks can approach the desired output behavior very effectively and also maintain high utilization of the network through the proposed self-tuning fuzzy congestion control scheme. Low cell loss ratio and high link utilization are achieved.