基于CEC2014基准测试集的和声搜索算法研究  

A Study of Harmony Search Algorithm Based CEC2014 Benchmark Functions

作　　者：赵瑞壮 李泓进 林初欣 陈泽世 可汗 Zhao Ruizhuang;Li Hongjin;Lin Chuxin;Chen Zeshi;Ke Han(School of Science,Jiangxi University of Science and Technology,Ganzhou,Jiangxi 341000,China)

机构地区：[1]江西理工大学理学院,江西赣州341000

出　　处：《机电工程技术》2025年第1期86-91,共6页Mechanical & Electrical Engineering Technology

基　　金：江西省研究生教育创新计划项目(YC2022-S692)。

摘　　要：作为一种基于群体的智能优化算法,和声搜索算法(Harmony Search,HS)因其具有控制参数少、健壮性强、结构简单等优点,引起了众多研究人员的广泛关注。然而随着研究工作的不断深入,和声搜索算法存在过早收敛、容易陷入局部最优的不足。因此,国内外许多研究人员提出了一系列的改进和声搜索算法。按照改进思路从改进控制参数的设置方法、改进HS的搜索算子与结合其他算法的优化方法3个方面,对现有的改进和声搜索算法进行归纳总结。此外,基于CEC2014基准测试集设计对比实验来比较5种具有代表性和声搜索算法的性能表现。根据实验结果,分析了5种具有代表性和声搜索算法在单峰、多峰、混合和组合函数上的性能表现,并给出了5种具有代表性和声搜索算法在不同函数类型工程优化问题上的建议。As a group-based intelligent optimization algorithm,the harmony search(HS)algorithm attracts continuous attention from many researchers due to its advantages of fewer control parameters,strong robustness,and simple structure,etc.However,with the continuous deepening of the research work,HS has the weaknesses of premature convergence and easily falling into the local optimum.To solve the drawbacks of HS,many researchers from home and abroad propose a series of improved HSs.According to the improvement ideas,the improved HSs can be summarized into 3 types,which include improvements of setting control parameters,the search operator,and combining with other algorithms.In addition,a comparison experiment is designed on the CEC2014 benchmark functions to compare the performance of five representative HSs.Afterwards,the performances of five representative HSs on unimodal,multimodal,hybrid,and composition functions are analyzed based on the experimental results.Suggestions for five representative HSs in engineering optimization problems with different function types are provided.

关 键 词：和声搜索算法 参数策略 搜索算子 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]