Faster and More Stabile-Asymmetric Bidirectional Weighting Topologies on PSO
Tang Pu-ying, Fang Jun · 2006
A new algorithm, called the asymmetric fully informed particle swarm optimization (AFIPSO), is proposed. It mends the FIPSO algorithm by newly constructing weighting functions, which realizes asymmetric influence between two particles. Then six weighting functions and their four cross combinations are tested. The simulation results have shown that the performance of the algorithm with the compounding-form of weighting functions is far better than the primary version, faster and more stabile in convergence, especially with perfect convergent rate