Novel self-adaptive particle swarm optimization methods

Abstract

This paper proposes adaptive versions of the particle swarm optimization algorithm (PSO). These new algorithms present self-adaptive inertia weight and time-varying adaptive swarm topology techniques. The objective of these new approaches is to avoid premature convergence by executing the exploration and exploitation stages simultaneously. Although proposed PSOs are fundamentally based on commonly utilized swarm behaviors of swarming creatures, the novelty is that the whole swarm may divide into many sub-swarms in order to find a good source of food or to flee from predators. This behavior allows the particles to disperse through the search space (diversification) and the sub-swarm, where the worst performance dies out while that with the best performance grows by producing offspring. The tendency of an individual particle to avoid collision with other particles by means of simple neighborhood rules is retained in these algorithms. Numerical experiments show that the new approaches, survival sub-swarms adaptive PSO (SSS-APSO) and survival sub-swarms adaptive PSO with velocity-line bouncing (SSS-APSO-vb), outperform other competitive algorithms by providing the best solutions on a suite of standard test problem with a much higher consistency than the algorithms compared.