Skip Neighborhood Hybrid Particle Swarm Optimization Algorithm

Abstract

Traditional Particle Swarm Optimization (PSO) uses single search strategy and is difficult to balance the global search with local search, and easy to fall into local optimization, a new algorithm which integrates global search with local neighborhood search is presented. The algorithm performs the global search in parallel with the local search by the feedback of the global optimal particle and the information interaction of local neighborhood. Meanwhile, with a new neighborhood topology to control the search space, the algorithm can avoid the local optimization successfully. Tested by four classical functions, the new algorithm performs well on optimization speed, accuracy and success rate.