Small world network based dynamic topology for particle swarm optimization

A new particle optimization algorithm with dynamic topology is proposed based on `small world´ network. The technique imitates the dissemination of information in a `small world network´ by dynamically updating the neighborhood topology of particle swarm optimization. The proposed dynamic neighborhood strategy can effectively coordinate the exploration and exploitation ability of particle swarm optimization. Simulations demonstrated that convergence of the swarms is guaranteed. Experiments demonstrated that the proposed method maintained the population diversity and enhanced the global search ability.