Particle Swarm Optimization based on Multiple Swarms and Opposition-based Learning*

Standard particle swarm optimization is easy to fall into local optimum and has the problem of low precision. To solve these problems, the paper proposes an effective approach, called particle swarm optimization based on multiple swarms and opposition-based learning, which divides swarm into two subswarms. The 1st sub-swarm employs PSO evolution model in order to hold the self-learning ability; the opposite solution of particle and the optimum between two sub-swarms are introduced into the 2nd sub-swarm which adopts new evolution model with boosting self-escaping and society learning ability of particle. Our method can improve the diversity of swarm and the ability of exchanging information. Experiments are conducted on a set of well-known benchmark functions to verify the performance of MSOL-PSO, comparing with other PSO variants; the results demonstrate promising performance of our approach on stability and efficiency. Keywords—particle swarm optimization; evolutionary model; multiple swarms; opposition-based learning;