融合入侵杂草算子的蝙蝠算法  

BAT ALGORITHM FUSED WITH INVASIVE WEED OPERATOR

作　　者：屈迟文[1] 傅彦铭[2] 侯勇顺 

机构地区：[1]百色学院数学与计算机信息工程系,广西百色533000 [2]广西大学计算机与电子信息学院,广西南宁530004 [3]重庆邮电大学计算机科学与技术学院,重庆400065

出　　处：《计算机应用与软件》2015年第4期243-246,275,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(61262072);广西高校科研项目(2013YB247)

摘　　要：针对基本蝙蝠算法在寻优后期存在寻优精度低、早熟及易陷入局部最优等缺陷,提出一种融合入侵杂草生长繁殖局部搜索的蝙蝠算法。该算法将入侵杂草算法的杂草生长繁殖、空间扩散和竞争机制融合到蝙蝠算法中,采用动态调整杂草空间扩散算子的标准差,使算法既增加了全局搜索能力,同时也提高了局部搜索能力。通过对8个标准测试函数的仿真测试,结果表明该算法具有较优的全局搜索和局部搜索能力,在寻优精度和收敛性方面都比基本蝙蝠算法有较大的提高。In its later period of optimisation,original bat algorithm has the defects of low accuracy,prematurity and easy to fall into local optimum,etc.To overcome them,we present such a bat algorithm which fuses invasive weed growth and reproduction.It fuses the weed growth and reproduction,spatial diffusion and competition mechanism of invasive weeds algorithm into bat algorithm and dynamically adjusts the standard deviation of weeds spatial diffusion operator,thus the algorithm increases the global and local search abilities both.Through the simulation testing on 8 standard test functions,the results show that this algorithm has better global and local search abilities,in the aspects of optimisation accuracy and convergence it gains greater improvement than the basic bat algorithm.

关 键 词：蝙蝠算法 生长繁殖算子 空间扩散算子 寻优精度 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]