An improved particle swarm optimization algorithm for global numerical optimization

Abstract

This paper presents an improved particle swarm optimization algorithm (IPSO) for global numerical optimization. The IPSO uses more particles' information to control the mutation operation. A new adaptive strategy for choosing parameters is also proposed to assure convergence of the IPSO. Meanwhile, we execute the IPSO to solve eight benchmark problems. The results show that the IPSO is superior to some existing methods for finding the best solution, in terms of both solution quality and algorithm robustness. © Springer-Verlag Berlin Heidelberg 2006.