一种新的自适应步长果蝇优化算法  

Research of the Self-adaptive Step Fruit Fly Optimization Algorithm

作　　者：段艳明[1] 肖辉辉[1,2] 

机构地区：[1]河池学院计算机与信息工程学院,江西宜州546300 [2]江西财经大学信息管理学院,南昌330013

出　　处：《河南师范大学学报（自然科学版）》2016年第1期161-168,共8页Journal of Henan Normal University(Natural Science Edition)

基　　金：国家自然科学基金(61173146);广西高校科研项目(KY2015LX332;KY2015LX334);校级项目(XJ2015QN003);江西省研究生创新项目(YC2015-B054);河池学院计算机网络与软件新技术重点实验室项目(院科研2013(3)号)

摘　　要：针对基本果蝇优化算法(FOA)易陷入局部最优、寻优精度低和后期收敛速度慢的问题,提出了一种自适应步长果蝇优化算法(ASFOA).该算法在运行过程中根据上一代最优味道浓度判断值和当前迭代次数来自适应调整进化移动步长,使算法在初期的步长大而避免种群个体陷入局部最优,到后期果蝇移动的步长变小而获得更高的收敛精度解,并加快收敛速度.通过6个标准测试函数对改进算法进行仿真测试,结果表明ASFOA算法具有更好的全局搜索能力,其收敛精度、收敛速度均比FOA算法及参考文献中其他改进果蝇优化算法有较大的提高.According to the problem that fruit fly optimization algorithm has low convergence accuracy, slow conver- gence velocity and easily falling into local optimization, we present a self-adaptive step fruit fly optimization algorithm（AS- FOA）. ASFOA can adjust adaptively the moving step according to the optimal flavor concentration values and the number of it- erations during the evolution. The large step of ASFOA in the initial state ensure that the solution cannot be trapped into local optimum. While the small step of ASFOA in the later stage improves the convergence accuracy and computational efficiency. The simulation results of 6 standard benchmark functions show that the ASFOA algorithm has the advantages of better global searching ability, the improved algorithm is much better than basic FOA,FOAAM and ACFOA in the respects of convergence precision convergence speed

关 键 词：自适应 果蝇优化算法 收敛速度 味道浓度 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]