一种全局收敛的PSO算法及其收敛分析

A global convergence algorithm of particle swarm optimization and its convergence analysis

在分析了速度因子对微粒群算法影响的基础上,针对以往算法的弱点,提出了一种基于Gaussian变异全局收敛的粒子群算法.该算法使用全局变异因子使粒子具有了良好的全局搜索能力,并证明了它能以概率1收敛到全局最优解.同时使用了局部变异因子,使算法在局部搜索过程中具有较高的搜索精度.典型函数优化的仿真结果表明,该算法具有寻优能力强、搜索精度高、稳定性好等优点,适合于工程应用中的函数优化问题.

Based on the analysis of the mutation operator having impacts on the particle swarm optimization, a global Gaussian particle swarm optimization （GGPSO） is proposed to overcome the problem of previous algorithm. A global mutation operator is used to make the particle have excellent ability of search in a global scope. Meanwhile this algorithm can converge to the global optimization solution with probability one. Furthermore, for improving the searching ability in local area, the algorithm uses the local mutation operator to make the algorithm behave well in local searching. Experiment simulations show that the proposed algorithm has powerful optimizing ability, good stability and higher optimizing precision, so it can be applied in optimization problems.