An accelerate convergence particle swarm optimization algorithm. (Chinese. English summary) Zbl 1240.65189

Summary: An accelerate convergence particle swarm optimization (ACPSO) algorithm is proposed based on analyzing the convergence of basal particle swarm optimization (BPSO) algorithm. The convergence speed of the ACPSO algorithm is very quickly through theoretical analysis. Then, the parameters in this algorithm are optimized. The mutation operator of the depending on segmental worst particles’ information is shown to escape a local optimum. The performance of the ACPSO algorithm with the optimal parameters is tested on several classical functions by comparing it with four classical PSO algorithms. The experimental results show that the ACPSO algorithm is efficient and robust. Especially, the convergence speed of ACPSO is superior to several classical PSO algorithms.


65K05 Numerical mathematical programming methods
90C15 Stochastic programming