Recent advances in particle swarm optimization via population structuring and individual behavior control

Particle swarm optimization (PSO) is an import bionic algorithm, inspired by the behaviors of gregarious colony such as bees, birds and fish. Since PSO was proposed in 1995 as a kind of swarm intelligence, many improved versions have been developed from different angles. In a swarm, population structures and individual behavior are the key elements for it to evolve. Therefore, in this paper we classify recent PSOs according to their development. Population structures are the foundation of a swarm. Thus some developments are discussed in accordance with the classification of single population and multiple sub-populations. Then the researches on static and dynamic topologies are also reviewed. After that, the improvements on individual behavior control are shown. Finally, some research directions to advance PSO are pointed out.