PSO方法的收敛性及基于微分演化的参数确定策略

Convergence property and parameters-selection strategy base on differential evolution of PSO

对已有关于PSO收敛性的研究结果进行了必要的修正和完善,并提出了一种不依赖个人经验的参数选择策略。针对特定问题,将PSO方法的性能表示成参数的函数,从而将参数选择问题转变成函数优化问题。同时,采用微分演化方法来确定PSO的最佳参数,收到了较好的效果。

It was modified and completed that the previous research result on PSO convergence property and proposed a new strategy on parameter selection which did not depend on expert experience. It transformed the parameter-selection problem into functional optimization problem by creating a function of the PSO property parameters. The result is also very promising in finding the optimal parameters of PSO by differential evolution.