Particle swarm optimization with chaotic velocity clamping (CVC-PSO)

This article proposes a novel approach in particle swarm optimization (PSO) that combines chaos and velocity clamping with the aim of eliminating its known disadvantage that enforces particles to continue searching in search space boundaries. This problem reduces the performance of algorithm in obtaining the global optimum. This heuristic approach is called PSO with chaotic velocity clamping (CVC-PSO). Chaos is the study of non-linear dynamic systems which have extreme sensitivity to initial conditions called butterfly effect. In this paper, we use logistic equation to generate fully chaotic and randomness sequences in order to provide the global exploration possibility for particles which have left the search space because of having large velocity update. Finally, the experimentally obtained results of CVC-PSO and other algorithms, such as genetic algorithm (GA), standard PSO with inertia weight and improved imperialist competitive algorithm (CICA), are listed in different tables for comparison. The obtained results represent the success of CVC-PSO against other algorithms.