融合学习差异与Lévy飞行的动态平衡正余弦算法
Dynamic balance sine cosine algorithm combining learning difference with Lévy flight
李聪 1刘昊 1赵雨微1
作者信息
- 1. 辽宁科技大学 理学院,辽宁 鞍山 114051
- 折叠
摘要
为了提升正余弦算法的收敛性能,本文提出一种融合学习差异与Lévy飞行的动态平衡正余弦改进算法,定义为SCALLD算法.通过引入学习差异策略,减少搜索个体对其位置信息的依赖,增强全局探索能力;加入Lévy飞行机制,丰富种群多样性,提升探索能力;采用动态平衡策略,平衡探索与开发能力,提高收敛速度和稳定性.在CEC2022基准测试函数上的实验表明,与六种算法相比,SCALLD展现出更优的收敛性能和稳定性,Wilcoxon秩和检验进一步证明了SCALLD的竞争优势,为解决复杂优化问题提供参考.
Abstract
In order to improve the convergence performance of sine cosine algorithm,this paper proposes an improved dynamic balance sine cosine algorithm that integrates learning difference and Lévy flight,which is defined as SCALLD algorithm.By introducing the learning difference strategy,the search individu'l's depen-dence on its location information is reduced and the global exploration ability is enhanced.Adding Lévy flight mechanism to enrich population diversity and improve exploration ability;Adopting the dynamic balance strat-egy to balance exploration and exploitation capabilities,to improve convergence speed and stability.Experi-ments on the CEC2022 benchmark functions show that SCALLD demonstrates superior convergence perfor-mance and stability compared to the six comparison algorithms.Wilcoxon rank sum test further proves SCALLD's competitive advantage and provides a reference for solving complex optimization problems.
关键词
正余弦算法/智能优化算法/学习差异策略/Lévy飞行/动态平衡Key words
sine cosine algorithm/intelligent optimization algorithm/learning difference strategy/Lévy flight/dynamic balance引用本文复制引用
出版年
2024
NETL
NSTL
万方数据