动态适应布谷鸟搜索算法 - 控制与决策.PDFVIP

第29 卷 第4 期 控 制 与 决 策 2014 年 4 月 Vol. 29 No. 4 Control and Decision Apr. 2014 文章编号: 1001-0920 (2014) 04-0617-06 DOI: 10.13195/j.kzyjc.2012.1815 动态适应布谷鸟搜索算法 张永韡, 汪 镭, 吴启迪 ( 同济大学电子与信息工程学院，上海201804) ´ 摘 要: 介绍一种新的生物启发算法—–布谷鸟搜索(CS) 及其相关的Levy 飞行搜索机制. 为了进一步提高算法的适 应性, 将反馈引入算法框架, 建立了CS 算法参数的闭环控制系统. 将Rechenberg 的1/5 法则作为进化的评价指标, 引 入学习因子平衡种群的多样性和集中性, 提出动态适应布谷鸟算法(DACS). 最后, 通过数值实验验证了所提出算法 的有效性. ´ 关键词: 布谷鸟算法；Levy飞行；动态适应；反馈控制 中图分类号: TP399 文献标志码: A Dynamic adaptation cuckoo search algorithm ZHANG Yong-wei, WANG Lei, WU Qi-di (School of Electronics and Information Engineering，Tongji University ，Shanghai 201804，China ．Correspondent: ZHANG Yong-wei ，E-mail ：yongwzhang@) ´ Abstract: A novel bio-inspired algorithm, cuckoo search(CS), is introduced along with the related Levy flight mechanism. In order to improve the adaptation of this algorithm, a feedback control scheme of algorithm parameters is adopted in CS. By utilizing Rechenberg’s 1/5 criteria to evaluate evolution process, and introducing the learning factor, the diversification and intensification of population are well balanced. The dynamic adaptation cuckoo search(DACS) algorithm is proposed. Finally, numerical experiment results show the effectiveness of the proposed algorithm. ´ Key words: cuckoo search ；Levy flight ；dynamic adaptation ；feedback control 0 引引引 言言言 算法应用于前馈神经网络训练; 文献[6] 将一种混合 [1]