An analysis of roulette selection in early particle swarm optimizing

In this paper, we present a novel particle swarm optimizer combined with the roulette selection operator. The modified algorithm provides a mechanism to restrain super particles in early stage and can effectively avoid the premature problem. It is empirically tested and compared with other published methods on several famous benchmark functions. The computational results illustrate that the proposed algorithm has the potential to achieve higher success ratio and better solution quality, especially for multimodal function optimization