Biomimicry of parasitic behavior in a coevolutionary particle swarm optimization algorithm for global optimization

Abstract

The declining of population diversity is often considered as the primary reason for solutions falling into the local optima in particle swarm optimization (PSO). Inspired by the phenomenon that parasitic behavior is beneficial to the natural ecosystem for the promotion of its biodiversity, this paper presents a novel coevolutionary particle swarm optimizer with parasitic behavior (PSOPB). The population of PSOPB consists of two swarms, which are host swarm and parasite swarm. The characteristics of parasitic behavior are mimicked from three aspects: the parasites getting nourishments from the host, the host immunity, and the evolution of the parasites. With a predefined probability, which reflects the characteristic of the facultative parasitic behavior, the two swarms exchange particles according to particles’ sorted fitness values in each swarm. The host immunity is mimicked through two ways: the number of exchange particles is linearly decreased over iterations, and particles in the host swarm can learn from the global best position in the parasite swarm. Two mutation operators are utilized to simulate two aspects of the evolution of the parasites in PSOPB. In order to embody the law of “survival of the fittest” in biological evolution, the particles with poor fitness in the host swarm are removed and replaced by the same numbers of randomly initialized particles. The proposed algorithm is experimentally validated on a set of 21 benchmark functions. The experimental results show that PSOPB performs better than other eight popular PSO variants in terms of solution accuracy and convergence speed.