An Improved Particle Swarm Optimization for Continuous Problems

摘要

This paper describes an improved particle swarm optimization (PSO) algorithm that combines stochastic local search (SLS) heuristics,named PSOSLS, to solve costly procedure of search and premature convergence for continuous function optimization problems. The SLS is embedded in the PSO to improve the proposed heuristics. During the global search process, our algorithm can enhance the local search ability of particle swarm optimization thought adding random perturbation to local search. Some optimization tests on many different benchmark optimization problems show that PSOSLS can search for global optima in difficult multimodal optimization problems and reach better solutions than original PSO algorithm.