北京航空航天大学学报 ›› 2021, Vol. 47 ›› Issue (12): 2579-2593.doi: 10.13700/j.bh.1001-5965.2020.0322

• 论文 • 上一篇    下一篇

一种改进X-best引导个体和动态等级更新机制的鸡群算法

张可为1,2, 赵晓林1, 何利3, 李宗哲1   

  1. 1. 空军工程大学 装备管理与无人机工程学院, 西安 710051;
    2. 空军工程大学 研究生院, 西安 710051;
    3. 中国人民解放军西安飞行学院, 西安 710306
  • 收稿日期:2020-07-07 发布日期:2022-01-06
  • 通讯作者: 李宗哲 E-mail:lzz144@163.com
  • 基金资助:
    国家自然科学基金(61503405)

A chicken swarm optimization algorithm based on improved X-best guided individual and dynamic hierarchy update mechanism

ZHANG Kewei1,2, ZHAO Xiaolin1, HE Li3, LI Zongzhe1   

  1. 1. Equipment Management and UAV Engineering College, Air Force Engineering University, Xi’an 710051, China;
    2. Graduate School, Air Force Engineering University, Xi’an 710051, China;
    3. PLA Air Force Xi’an Flight Academy, Xi’an 710306, China
  • Received:2020-07-07 Published:2022-01-06
  • Supported by:
    National Natural Science Foundation of China (61503405)

摘要: 在群智能算法的改进中,常利用优秀个体加速算法收敛,但对其依赖过度会导致种群多样性和算法全局收敛性下降的现象。对此,提出一种改进X-best引导个体和动态等级更新机制的鸡群算法。首先,在个体更新阶段不仅引入优秀个体加速收敛,并且通过普通个体对优秀个体的影响进行适当平衡,因此,优秀个体与普通个体的信息都能得到利用,进而种群多样性和算法全局收敛性得到提升。其次,通过对等级更新参数进行动态优化,加强了种群等级更新机制对算法收敛的促进作用。最后,经过时间复杂度与收敛性分析,证明了改进算法仍具有简单性和全局收敛性。仿真结果表明:所提出的改进算法较其他对比算法在寻优精度、寻优成功率和收敛速度等方面都具有明显优势。

关键词: 鸡群算法, X-best引导, 动态等级更新, 收敛性分析, 函数优化

Abstract: In the improvement process of swarm intelligence algorithms, elite individuals are often used to accelerate the convergence, but excessive dependence on them will lead to the decline of population diversity and global convergence. In this regard, a chicken swarm optimization algorithm based on improved X-best guided individual and dynamic hierarchy update mechanism is proposed in this paper. Firstly, in the individual update stage, elite individuals are introduced into the search equation to accelerate the convergence, while the ordinary individuals are also introduced into the search equation to balance the influence of the elite individuals. Therefore, the information of elite and ordinary individuals can be fully used, and the population diversity and global convergence are improved. Secondly, by dynamically optimizing the hierarchy update parameter, the promotion effect of the population hierarchy update mechanism on the convergence is strengthened. Finally, through complexity and convergence analysis, the simplicity and global convergence of IDCSO are proved. The simulation results show that IDCSO has obvious advantages over other comparative algorithms in terms of optimization accuracy, optimization success rate and convergence speed.

Key words: chicken swarm optimization algorithm, X-best guided, dynamic hierarchy update, convergence analysis, function optimization

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发