改进万有引力搜索算法在函数优化中的应用  

Application of improved gravitational search algorithm in function optimization

作　　者：刘小刚 欧阳自根[2] LIU Xiao-gang;OUYANG Zi-gen(School of Science,Xijing University,Xi’an 710123,China;School of Mathematics and Physics,University of South China,Hengyang 421000,China)

机构地区：[1]西京学院理学院,西安710123 [2]南华大学数理学院,湖南衡阳421000

出　　处：《沈阳工业大学学报》2021年第2期193-197,共5页Journal of Shenyang University of Technology

基　　金：陕西省教育厅专项科研计划项目(16JK2213).

摘　　要：为了克服标准的万有引力搜索算法在函数优化中迭代速度慢、易陷入局部最优等问题,基于加强算法的性能,研究了新的策略.结合粒子群算法的开采能力和万有引力搜索算法的勘察能力,得到了基本粒子群万有引力搜索混合算法.对混合算法中的加速因子进行改进并引入了动量因子,提出了一种改进的粒子群万有引力搜索混合算法(IPSOGSA).结果表明:与粒子群算法、万有引力搜索算法、基本粒子群万有引力搜索混合算法相比,改进的粒子群万有引力搜索混合算法在非线性的复杂函数优化中具有更好的寻优能力.In order to overcome the problems,e.g.lower iteration speed and liability of local optimum,caused by standard gravitational search algorithm in function optimization,a novel strategy based on strengthening the performance of a newly proposed algorithm was studied.Combining the exploitation ability of particle swarm optimization(PSO)algorithm and the reconnaissance ability of gravitational search algorithm(GSA),a basic hybrid particle swarm optimization gravitational search algorithm(PSOGSA)was obtained.In addition,the acceleration factors in this hybrid algorithm were modified,a momentum factor was employed,and an improved particle swarm optimization gravitational search algorithm(IPSOGSA)was proposed.The results show that IPSOGSA has better searching ability in nonlinear complex function optimization,compared with PSO,GSA and PSOGSA.

关 键 词：局部最优问题 万有引力搜索算法 粒子群算法 混合算法 加速因子 动量因子 测试函数 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]