摘要
针对布谷鸟优化算法存在的易陷入局部最优、收敛速度慢等问题,提出了融合正弦、余弦和种群初始化策略的布谷鸟算法。在布谷鸟初始种群产生时引用一种结合均匀化与随机化的策略,以半均匀半随机的方式产生初始种群,有效地减小了因为初始种群完全随机而产生的随机误差,提高了算法寻优速度;分别在全局搜索和局部搜索的上一代鸟窝位置处引入正弦、余弦算子,灵活调整上一代鸟窝位置,有效克服算法易陷入局部最优的缺陷,提高寻优搜索能力;引入了指数型动态概率代替固定概率,动态平衡全局搜索和局部搜索。与其他4种算法相比较并通过6个测试函数仿真,结果表明:融合正弦、余弦和种群初始化策略的布谷鸟算法收敛速度更快,求解精度更高,具有更好的寻优性能。
The cuckoo optimization algorithm is easy to fall into the local optimal,slow convergence and so on,a cuckoo algorithm combining sine,cosine and population initialization strategy was proposed.In the process of initial population generation of cuckoo,a strategy combining homogenization and randomization was introduced,and the initial population was generated in a semi-uniform and semi-random way,which effectively reduces the random error caused by the completely random initial population,and improves optimization speed of the algorithm;Sine and cosine operators are introduced in the bird′s nest position of the previous generation of global search and local search respectively to flexibly adjust the previous generation bird′s nest position,which effectively overcome the defect that the algorithm is easy to fall into local optimal,and improve the ability to search for optimization.The exponential dynamic probability is introduced to replace the fixed probability,and the global search and local search are balanced dynamically.Compared with the four algorithms and simulated by six test functions,the results show that the cuckoo algorithm combining sine,cosine and population initialization strategy has faster convergence speed,higher solving accuracy and better optimization performance.
作者
张珍珍
贺兴时
于青林
杨新社
ZHANG Zhenzhen;HE Xingshi;YU Qinglin;YANG Xinshe(School of Science, Xi’an Polytechnic University, Xi’an 710048, China;School of Mathematics and Statistics, Thompson Rivers University, Kamloops V2C0C8, Canada;School of Science and Technology, Middlesex University, London NW4 4BT, UK)
出处
《纺织高校基础科学学报》
CAS
2021年第4期102-109,共8页
Basic Sciences Journal of Textile Universities
基金
国家自然科学基金(12001417)
陕西省智慧医疗评价指标体系构建及评价模型研究(2019KPM141)。
关键词
布谷鸟算法
初始化
均匀化与随机化
正余弦算子
动态概率
cuckoo algorithm
initialization
homogenization and randomization
sine and cosine operator
dynamic probability
