A Bi-swarm Particle Swarm Optimization with Cooperative Co-evolution

In order to develop the global performance of particle swarm optimization (PSO), the paper proposes a bi-swarm particle swarm optimization with cooperative co-evolution (BPSO-CC). BPSO-CC adopts two swarms to go on the search, the first swarm is designated to conduct the coarse search in the whole space, while the second swarm is generated periodically surround the first swarm and designated to make the fine search in the local search area around the first swarm. With the same aim to find out the global optimum, the two swarms go on their search in parallel. at the same time, they keep the cooperative co-evolution through sharing and exchanging the valid information, which can make them search in correct direct and improve the convergent efficiency of PSO validly. The proposed BPSO-CC has been applied to solve some benchmark functions with large scales, the simulation results demonstrate that BPSO-CC is a robust technique for complex optimizations and performs better than SPSO, not only in the convergence precision but also in the efficiency.