电子学报 ›› 2019, Vol. 47 ›› Issue (2): 282-288.DOI: 10.3969/j.issn.0372-2112.2019.02.004

• 学术论文 • 上一篇    下一篇

均衡单进化布谷鸟算法

傅文渊   

  1. 1. 华侨大学信息科学与工程学院, 福建厦门 361002;
    2. 中山大学电子与信息工程学院, 广东广州 510006;
    3. 厦门市专用电路系统重点实验室, 福建厦门 361008;
    4. 福建省电机控制与系统优化调度工程技术研究中心, 福建厦门 361002
  • 收稿日期:2018-08-08 修回日期:2018-10-09 出版日期:2019-02-25
    • 作者简介:
    • 傅文渊 男,1982年9月出生,福建邵武人,讲师,主要从事智能信号优化与智能学习控制、电路与系统设计及嵌入式系统设计等方面的研究工作.E-mail:fwy@hqu.edu.cn
    • 基金资助:
    • 国家自然科学基金 (No.61203369); 福建省自然科学基金 (No.2015J1263); 福建省中青年教育科研 (No.JA15037)

Equilibrium Single Evolution Based Cuckoo Search Algorithm

FU Wen-yuan   

  1. 1. College of Information Science and Engineering, Huaqiao Univesity, Xiamen, Fujian 361002, China;
    2. School of Electronics and Information Technology, Sun Yat-sen University, Guangzhou, Guangdong 510006, China;
    3. Xiamen Key Laboratory of ASIC System, Xiamen, Fujian 361008, China;
    4. Fujian Engineering Research Center of Motor Control and System Optimal Schedule, Xiamen, Fujian 361002, China
  • Received:2018-08-08 Revised:2018-10-09 Online:2019-02-25 Published:2019-02-25
    • Supported by:
    • National Natural Science Foundation of China (No.61203369); Natural Science Foundation of Fujian Province,  China (No.2015J1263); Fujian Province Young and Middle-aged Teacher Education Research Project (No.JA15037)

摘要: 针对布谷鸟算法采用整体评价策略处理多维度自变量相关优化问题时,维度耦合现象会恶化算法的搜索速度和收敛精度,提出均衡单进化的布谷鸟算法(ESCES).该算法给出一种新型的均衡单进化函数评价策略,即每一代进化只随机更新目标函数的单个维度,并且随机更新的维度服从均匀分布,避免多维度之间互相干扰.同时,提出两种新型随机游动步长更新学习律,提高了优化算法的全局搜索速度和收敛精度.实验测试结果和显著性统计结果表明,ESCES算法与5个改进CS算法及7个其它最新智能优化算法相比,在全局寻优性能、搜索速度和收敛精度上均获得较大的改进.

关键词: 进化, 评价策略, 布谷鸟算法, 发现概率

Abstract: For the whole evaluation strategy in cuckoo search algorithm in the face of multi-dimension function optimization problems,the coupling phenomena among dimensions will deteriorate the search speed and convergence accuracy.Therefore,a new cuckoo search algorithm based on the equilibrium single evolution mechanism is proposed.Then,a new equilibrium single evolution evaluation strategy is also used to update randomly the single dimension of the objective function on each iteration.Note that the randomly updated dimensions obey the uniform distribution to avoid mutual interference between dimensions.Furthermore,two new random walking update laws are proposed to improve the global search speed and convergence accuracy.The results of the 10 benchmark functions and statistical significance demonstrate that ESCES algorithm has a great improvement in global optimization performance,search speed and convergence accuracy compared with the five modified CS algorithms and seven other state-of-the art algorithms.

Key words: evolution, evaluation strategy, cuckoo search algorithm, discovery probability

中图分类号: