Congestion control in computer networks: Application of piece-wise affine controller and particle swarm optimization

Congestion control plays a significant role in maintaining sufficient network throughput. Variety of methods and algorithms are proposed to solve the bottleneck issue. This paper describes a method based on a particle swarm optimization algorithm and piecewise affine controller for non-stationary, discrete, dynamical model of data exchange network. This solution allows active network nodes buffer utilization to be adapted to sudden unexpected network state changes. That leads to alleviate network congestion effects.