An Improved Multi-particle Swarm Co-evolution Algorithm

Illumined by phenomenon of co-evolution in nature, particle swarm optimization is combined with co-evolution, in this paper, an improved multi-particle swarm co-evolution algorithm is presented. In the process of evolution, particles not only have to exchange information with the others that are in its own sub-swarm but also are influenced by the particles from the other sub-swarms. By doing experiments on three benchmark functions, the results show that the algorithm avoids trapping into local optimum in certain extents and improves the precision of convergence.