Abstract
Particle Swarm Optimization (PSO) is an optimization method inspired by the flock behavior. In the original PSO, homogeneous particles search solutions. Several extensions where respective particles can have different search strategies have been proposed. In Heterogeneous PSO (HPSO), respective particles select their own search strategies from a strategy pool, which consists of five kinds of strategies. If the personal best value of a particle has not been improved for some iterations, the particle reselects its search strategy. The global search can be performed by the heterogeneity of search strategies. In Predator Prey Optimizer (PPO) is the PSO to which the predator-prey relationship has been introduced. Prey particles have to keep away from predator particles. Escape from local optima can be performed by the interaction of the two kinds of particles. In this paper, we introduce the predator-prey relationship into the search strategy pool of HPSO. We examine the search performance of our proposed methods and the effect of the diversification of search strategies.
