一种改进的粒子群优化算法

An Improved Particle Swarm Optimization Algorithm

针对粒子群优化算法早熟收敛现象,提出了一种改进的粒子群优化算法.该算法将模拟退火算法的"上山性"引入粒子群算法中,同时为了增加种群的多样性,将交叉和变异算子也结合进去,形成了一种新的改进粒子群算法.比较了高斯变异和柯西变异这两种变异算子对改进算法的影响.改进算法对典型函数的优化计算结果表明,与基本粒子群算法相比,改进算法能够更加快速有效的收敛到全局最优解,而且采用柯西变异算子的改进算法的效果比采用高斯变异算子的效果要好.

To solve the premature convergence problem of the Particle Swarm Optimization （PSO）, an improved PSO method was proposed. In the improved method, the ＂uphill＂ movement of SA was introduced and the operations of crossover and mutation was used to keep the diversity of the population. The effects of Gauss mutation operator and the Cauehy mutation operator on the improved algorithm were compared. The optimal results on benchmark functions demonstrate that the proposed method can get the global optimal result more quicky and more efficiently than the basic PSO, and the improved algorithm which use Cauchy mutation operator outperforms the algorithm which use Guass mutation operator regarding the quality of solutions on benchmark functions.