Abstract
A new particle optimization algorithm with dynamic topology is proposed based on
a small world network. The technique imitates the dissemination of information in a
small world network by dynamically updating the neighborhood topology of the particle
swarm optimization(PSO). In comparison with other four classic topologies and
two PSO algorithms based on small world network, the proposed dynamic neighborhood
strategy is more eÆective in coordinating the exploration and exploitation ability
of PSO. Simulations demonstrated that the convergence of the swarms is faster than
its competitors. Meanwhile, the proposed method maintains population diversity and
enhances the global search ability for a series of benchmark problems.