Genetic-sigmoid random early detection covariance control as a jitter controller

Abstract

The congestion control problem of the intermediate nodes in the Internet has received extensive attention in control community. In this study, a novel intelligent sigmoid random early detection controller based on a genetic algorithm is presented to solve the problem of active queue management. In this work, an optimisation theoretic method is used to achieve a better performance in queue management in transmission control protocol networks. A covariance control based on the proposed method is implemented to stabilise and control the fluctuation of the queue length, which in turn results in minimum jitter. The authors illustrate their results with NS2 simulations.