改进的万有引力搜索算法  

Improved Gravitational Search Algorithm

作　　者：张秀玲[1,2] 臧佳音[1] 樊红敏[1] 赵亮[1] 

机构地区：[1]燕山大学河北省工业计算机控制工程重点实验室,河北秦皇岛066004 [2]国家冷轧板带装备及工艺工程技术研究中心,秦皇岛066004

出　　处：《沈阳大学学报（自然科学版）》2014年第6期468-472,共5页Journal of Shenyang University：Natural Science

基　　金：国家自然科学基金资助项目(61074099)

摘　　要：针对万有引力搜索算法对一些优化问题的搜索精度不高,容易出现早熟的问题,提出了改进的万有引力搜索算法.该算法借鉴生物界中雁群的飞行特征和加权平均法,扩大了搜索范围,加强了粒子间的合作与竞争.通过对6个基准函数的仿真测试,MATLAB仿真结果表明新算法能更有效地提高全局搜索能力.Aiming at the problems that gravitational search algorithm easily falls into premature convergence and has bad performance in search accuracy,improved gravitational search algorithm（IGSA）is put forward.This new method,combining the flight characteristics of wild geese flock and the method of weighted mean,expands the search scope and strengthens the cooperation and competition between the particles.The simulation test of the 6benchmark functions by MATLAB confirms that the improved algorithm efficiently improves the global search ability.

关 键 词：万有引力搜索算法 雁群飞行特征 加权平均法 数值函数优化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]