一种参数动态更新的布谷鸟搜索算法  

New Cuckoo Search Algorithm with Parameter Dynamic Updating

作　　者：闫河 谢敏 赵其峰 李晓玲 YAN He;XIE Min;ZHAO Qi-feng;LI Xiao-ling(College of Liangjiang Artificial Intelligence,Chongqing University of Technology,Chongqing 401135,China;College of Computer Science and Engineering,Chongqing University of Technology,Chongqing 400054,China)

机构地区：[1]重庆理工大学两江人工智能学院,重庆401135 [2]重庆理工大学计算机科学与工程学院,重庆400054

出　　处：《小型微型计算机系统》2022年第9期1846-1852,共7页Journal of Chinese Computer Systems

基　　金：国家重点研发计划“智能机器人”重点专项项目(2018YFB1308602)资助;国家自然科学基金面上项目(61173184)资助;重庆市自然科学基金项目(cstc2018jcyjAX0694)资助.

摘　　要：为提升布谷鸟算法对高维问题空间的搜索能力及算法的稳定性,提出了一种参数动态更新的布谷鸟搜索算法.该算法首先选用柯西随机数,动态更新迭代学习的步长因子,并结合Lévy飞行策略生成新解;在新解生成之后,提出了一种正态扰动策略生成干扰解以增加解的搜索空间;对迭代中的新解和干扰解,采用模拟退火算法得到优势解,从而避免算法陷入局部最优,提升了算法的容差性;并提出一种轮盘赌选择和双向随机搜索策略强化迭代中优势解的学习.实验结果表明改进算法拥有较高的准确性和稳定性.In order to improve the search ability and stability of the cuckoo algorithm in high dimensional space,a new cuckoo search algorithm with parameter dynamic updating is proposed.In this paper,a new solution is calculated by Lévy flight strategy combing with step size,and the step size is dynamically updated by Cauchy random number during the iteration.A disturbance solution,calculating by Normal disturbance strategy,is to expand the search space of solution after calculating the new solution.Obtain a better solution between the new solution and the disturbed one by the simulated annealing algorithm during the iteration,which can avoid trapping into local optimum and improve the tolerance of the algorithm.Strengthen the learning of advantage solution in the iteration by roulette wheel selection method and bidirectional random search method.Experimental results reveal that the improved algorithm has a high accuracy and stability.

关 键 词：布谷鸟搜索算法 柯西随机数 正态扰动策略 模拟退火 轮盘赌选择 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]