一种万有引力优化算法及其收敛性分析

摘要

在分析信息共享机制对粒子群算法的影响及万有引力的基础上,针对粒子群早熟收敛和搜索能力差的问题,提出了一种基于万有引力的算法,质点之间的信息共享是由它们之间引力大小决定的,而不是受随机量影响的,并对算法的收敛性进行了分析.函数优化仿真结果显示该算法简单且能够快速寻找到最优点,在与粒子群比较的过程中发现,该算法速度快,也有着不错的寻优能力,适用于工程中的函数优化问题.%Based on the analysis of information sharing mechanism having impact on particle swarm optimization and universal gravitation, a universal gravitation optimization (UGO) was proposed to overcome the problems of previous one which premature convergence and poor search capability. The information sharing mechanism determined by universal gravitation a-mong particles instead of random impacts, and the convergence of UGO was analyzed. Experiment simulation showed that it' s brief and easily get optima, slight slower then PSO, but had good optimization ability, so it could be applied in optimization problems.