动态搜索协同进化的果蝇优化算法  

Dynamic Search and Cooperative Learning for Fruit Fly Optimization Algorithm

作　　者：张水平[1] 陈阳[1] 丁小军 

机构地区：[1]江西理工大学信息工程学院,江西赣州341000

出　　处：《小型微型计算机系统》2018年第1期48-52,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61662028)资助

摘　　要：针对基本果蝇优化算法(FOA)求解高维复杂问题时求解精度低,收敛速度慢等问题,提出一种动态搜索协同进化的果蝇优化算法(DCFOA).通过加入精英个体提高种群多样性并以一个线性递减的牵引因子诱导精英个体从算法初期就协同寻优,扩大其全局其搜索能力.当算法后期个体聚集度变大果蝇个体多样性变低,引入搜索空间压缩的搜寻策略,将目标问题的空间域动态变化为一种自适应步长,帮助算法跳出局部最优而进行深度寻优.对6个经典测试函数的实验证明,该算法可以有效避免早熟收敛,改善收敛速度,提高求解精度.In this paper,dynamic search and cooperative learning for Fruit fly optimization algorithm （DCFOA） is proposed to solvethe problem of low precision and low convergence speed when the basic Fruit fly optimization algorithm （FOA） is used to solve high-dimensional complex problems. By adding elite individuals to improve the diversity of the population and with a linear decline of thetraction factor to induce elite individuals from the initial stage of the algorithm to optimize, expand its global search capabilities. Whenthe degree of individual aggregation becomes high in the later stage of the algorithm, the searching space of the fruit fly is reduced,in-troduce the search space compression search strategy,the spatial dynamic change target problem is a kind of adaptive step to help thealgorithm to jump out of the local optimum. Experiments on six classical test functions show that this algorithm can effectively avoidthe premature convergence,improve the convergence speed and improve the accuracy.

关 键 词：果蝇优化算法 牵引因子 精英个体 动态压缩空间 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]