Statistical analysis for vortex particle swarm optimization

This paper presents the statistical analysis of vortex particle swarm optimization (VPSO) which is a boost algorithm based on self-propelled particle swarms. In order to avoid local minima, the optimization algorithm uses two separated behaviors: translational and dispersion. This idea mimics living organism strategies such as foraging and predator avoidance. The dispersion is given by vortex behavior (circular movements) to scape from local minima. Via suitable parameter configuration is possible to switch between translational (convergence) and circular movements (dispersion). Performance of the algorithm is studied via statistical analysis results using well-known test functions.