Roaming Behavior of Unconstrained Particles

Abstract

It has been shown recently that unconstrained particles that follow the position and velocity update rules of a standard global best particle swarm optimization algorithm leave the boundaries of the search space within the first few iterations of the search process. Provided that a better solution does not exist outside of the search boundaries, these roaming particles are eventually pulled back within the search boundaries. This article illustrates the consequence of roaming particles should better solutions exist outside of the search boundaries, namely that particles are pulled outside of the search boundaries and that such infeasible solutions are found. The article also evaluates the hypothesis that it is the roaming behavior of unconstrained particles that improves the ability of particle swarm algorithms to locate feasible solutions outside of the particle initialization space.