基于柯西变异的果蝇优化算法
|
摘要
针对果蝇优化算法是模仿果蝇寻找食物行为而进行全局搜索最优解的新算法,该算法存在容易陷入局部最优解和收敛速度慢的缺点。提出了一种基于柯西变异的果蝇优化算法,利用柯西分布具有较高的两翼概率特性从而容易产生一个远离原点的随机生成数,即柯西分布有一条很长的尾巴。所以在果蝇个体利用嗅觉搜索食物之随机方向距离上引入柯西变异算子代替原来随机方向变异算子进行扰动,从而容易跳出局部最优。最后通过数值仿真实验对6个标准测试函数来进行作对比检验,结果表明该算法在求解高维函数优化问题更好。
Based on the shortcomings of Fruit Fly Optimization Algorithm,which imitates the behavior of flies looking for food,such as low precision,slow convergence rate,and easily falling into local optimal solution,an improvement about the Fruit Fly Optimization Algorithm that is based on Cauchy mutation is put forward.Cauchy distribution has a high probability of the two wings characteristics,which can easily generate a randomly generated number away from the origin.That is,Cauchy distribution has a long tail.This method applies Cauchy mutation operator instead of the original direction of random mutation operator into the random directions of the individual fruit fly's sense of smell while looking for food,thus jumping out of local optimum.At last,a contrast test is conducted to compare the six standard test functions,with the results showing that this algorithm is much better in solving the high-dimensional function optimization problems.
Based on the shortcomings of Fruit Fly Optimization Algorithm,which imitates the behavior of flies looking for food,such as low precision,slow convergence rate,and easily falling into local optimal solution,an improvement about the Fruit Fly Optimization Algorithm that is based on Cauchy mutation is put forward.Cauchy distribution has a high probability of the two wings characteristics,which can easily generate a randomly generated number away from the origin.That is,Cauchy distribution has a long tail.This method applies Cauchy mutation operator instead of the original direction of random mutation operator into the random directions of the individual fruit fly's sense of smell while looking for food,thus jumping out of local optimum.At last,a contrast test is conducted to compare the six standard test functions,with the results showing that this algorithm is much better in solving the high-dimensional function optimization problems.
引文
[1]Pan W T.A new fruit fly optimization algorithm:taking the financial distress model as an example[J].knowledgebased Systems,2012,26:69-74.
[2]潘文超.果蝇最佳化演算法[M].台北,沧海书局,2011:10-12.
[3]潘文超.应用果蝇优化算法优化广义回归神经网络进行企业经营绩效评估[J].太原理工大学学报,2011,29(4):1-5.
[4]韩俊英,刘成忠,王联国.动态双子群协同进化果蝇优化算法[J].模式识别与人工智能,2013,26(11):1057-1067.
[5]韩俊英,刘成忠.基于细菌趋化的果蝇优化算法[J].计算机应用,2013,33(4):964-966.
[6]韩俊英,刘成忠.自适应变异的果蝇优化算法[J].计算机应用研究,2013,30(9):2641-2644.
[7]韩俊英,刘成忠.基于历史认知的果蝇优化算法[J].计算机科学与探索,2014,8(3):368-375.
[8]韩俊英,刘成忠.自适应调整参数的果蝇优化算法[J].计算机工程与应用,2014,,50(7):50-55.
[9]王雪刚,邹早建.基于果蝇优化算法的支持向量机参数优化在船舶操纵预报中的应用[J].上海交通大学学报,2013,47(6):884-888.
[10]马超,董玲.果蝇优化算法(FOA)步长改进及其多元函数最优化方法[J].专题研究,2013,26(13):90-93.
[11]程慧,刘成忠.基于混沌映射的混合果蝇优化算法[J].计算机工程,2013,39(5):218-221.
[12]侯越,赵贺,路小娟.基于萤火虫优化的BP神经网络算法研究[J].兰州交通大学学报,2013,32(6):24-27.
[13]胡能发.演化式果蝇算法及其应用研究[J].计算机技术与发展,2013,23(7):131-137.
[14]Holland J H.Oudine for a logical theory of adaptive systems[J].Joumal of the Association for Computing Machinery,1962,9(3):297-314.
[15]HOU,AND J H.Adaptation in natural and artificial systems[M].Ann Arbor,Mich:University of Michigan Press,1975.
[16]Kennedy J,Eberhart R.Particle$wal'nl optimization[C]//eroc0f t Ile IEEE Intemationnal Conference on Netural Networks,1995:1942-1948.[17]王战权,赵朝义,云庆夏.进化策略中基于柯西分布的变异算子改进探讨[J].系统工程,1997,17(4):49-54.[18]林川,冯全源.一种新的自适应粒子群优化算法[J].计算机工程,2008,34(7):181-183.
[2]潘文超.果蝇最佳化演算法[M].台北,沧海书局,2011:10-12.
[3]潘文超.应用果蝇优化算法优化广义回归神经网络进行企业经营绩效评估[J].太原理工大学学报,2011,29(4):1-5.
[4]韩俊英,刘成忠,王联国.动态双子群协同进化果蝇优化算法[J].模式识别与人工智能,2013,26(11):1057-1067.
[5]韩俊英,刘成忠.基于细菌趋化的果蝇优化算法[J].计算机应用,2013,33(4):964-966.
[6]韩俊英,刘成忠.自适应变异的果蝇优化算法[J].计算机应用研究,2013,30(9):2641-2644.
[7]韩俊英,刘成忠.基于历史认知的果蝇优化算法[J].计算机科学与探索,2014,8(3):368-375.
[8]韩俊英,刘成忠.自适应调整参数的果蝇优化算法[J].计算机工程与应用,2014,,50(7):50-55.
[9]王雪刚,邹早建.基于果蝇优化算法的支持向量机参数优化在船舶操纵预报中的应用[J].上海交通大学学报,2013,47(6):884-888.
[10]马超,董玲.果蝇优化算法(FOA)步长改进及其多元函数最优化方法[J].专题研究,2013,26(13):90-93.
[11]程慧,刘成忠.基于混沌映射的混合果蝇优化算法[J].计算机工程,2013,39(5):218-221.
[12]侯越,赵贺,路小娟.基于萤火虫优化的BP神经网络算法研究[J].兰州交通大学学报,2013,32(6):24-27.
[13]胡能发.演化式果蝇算法及其应用研究[J].计算机技术与发展,2013,23(7):131-137.
[14]Holland J H.Oudine for a logical theory of adaptive systems[J].Joumal of the Association for Computing Machinery,1962,9(3):297-314.
[15]HOU,AND J H.Adaptation in natural and artificial systems[M].Ann Arbor,Mich:University of Michigan Press,1975.
[16]Kennedy J,Eberhart R.Particle$wal'nl optimization[C]//eroc0f t Ile IEEE Intemationnal Conference on Netural Networks,1995:1942-1948.[17]王战权,赵朝义,云庆夏.进化策略中基于柯西分布的变异算子改进探讨[J].系统工程,1997,17(4):49-54.[18]林川,冯全源.一种新的自适应粒子群优化算法[J].计算机工程,2008,34(7):181-183.