改进的萤火虫算法及其工程应用  

Modified firefly algorithm and its engineering applications

作　　者：尹华一[1] 朱顺痣[1] 刘利钊[1] 张千宏[2] 

机构地区：[1]厦门理工学院计算机与信息工程学院,福建厦门361024 [2]贵州财经大学数学与统计学院,贵州贵阳550025

出　　处：《南京理工大学学报》2016年第6期653-660,共8页Journal of Nanjing University of Science and Technology

基　　金：国家自然科学基金(61503316;61373147;11361012);福建省科技计划对外合作项目(2016I0015);厦门市科技局高校对接企业项目(3502Z20153020);厦门理工学院高层次人才项目(YKJ13026R)

摘　　要：针对萤火虫算法存在易出现早熟收敛、后期收敛慢和精度低等问题,提出1种改进的萤火虫算法。采用反向学习策略对群体中个体位置进行初始化。引入Rosenbrock搜索以加快算法收敛和增强求解精度。对当前群体中最优萤火虫个体进行高斯混沌扰动以防止出现早熟收敛现象。选取6个标准函数进行仿真实验,并对2个标准工程应用问题进行求解。结果表明,该改进的萤火虫算法具有较强的全局优化性能。mA modified firefly algorithm is proposed aiming at the disadvantages of firefly algorithm such as premature convergence, slow convergence speed at a later stage and low solving precision. The individual location of groups is initialized by using the opposite learning strategy. Convergencespeed and solving precision are improved by using Rosenbrock search. Premature convergence is prevented by using a Gaussian chaos disturbance on the global optimum individual of each generation. Six standard functions are selected for simulation experiments, and two standard engineering applications are solved. The results show that this modified firefly algorithm has good global optimization performance.

关 键 词：萤火虫算法 反向学习 高斯混沌扰动 工程优化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]