Fuzzy logic target utilization and prediction for traffic control

Abstract

This paper presents a novel closed-loop traffic controller which is based on fuzzy set and fuzzy logic theory. The proposed controller has stable and robust operations under long round-trip delays and it does not use complicated analytical methods. A key feature of the controller is its ability to target link utilization dynamically based on the predicted future buffer condition in the switch. This is achieved by using fuzzy logic target utilization and fuzzy logic prediction. The prediction estimates the queue length one round-trip delay ahead. The predicted queue length, together with the queue growth rate and current queue length are used to produce a target utilization factor. The proposed fuzzy logic traffic control scheme is applied to both approximate and exact fair rate computation switch algorithms. The performance of fuzzy logic traffic control scheme is then compared with conventional schemes. Simulation results show that the fuzzy logic control scheme significantly reduces convergence time and overall buffer size requirements. It also exhibits lower queueing delay, delay variation and cell loss ratio.