Scholarlink

<p>This paper presents a memetic approach of Competitive Swarm Optimizer (CSO) named Self-Adaptive Competitive Swarm Optimizer (SACSO). We focus on two key parameters of CSO: scaling factor and population size and introduce a parameter sorting scheme and a linear population reduction strategy to enhance the performance of SACSO. Specifically, the parameter sorting scheme assigns smaller scaling factors to better-performing particle individuals to encourage exploitative searchability, while larger scaling factors are allocated to poor-performing particle individuals to promote exploration in unknown search areas. The linear population reduction strategy adapts to different optimization phases, emphasizing exploration in the early phase and enhancing exploitation in the later phase. To investigate the performance of SACSO, we conducted comprehensive numerical experiments on CEC2017, CEC2022, and seven classic engineering problems. The experimental results and statistical analysis demonstrate the competitiveness of SACSO against six state-of-the-art optimizers and four advanced CSO variants. Ablation studies further validate the contribution of the two introduced strategies. Finally, we applied SACSO to real-world Human-Powered Aircraft (HPA) design problems, and the results highlight the potential of SACSO for solving various real-world optimization challenges. The source code of SACSO is available at <a href="https://github.com/RuiZhong961230/SACSO">https://github.com/RuiZhong961230/SACSO</a>.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Self-adaptive competitive swarm optimizer: a memetic approach for global optimization and human-powered aircraft design

  • 期刊论文

  • 发表日期：

    2025年6月23日

查看全文
  • Rui Zhong,
  • Zhongmin Wang,
  • Ibrahim Al-Shourbaji,
  • Essam H. Houssein,
  • Pramod H. Kachare,
  • Abdoh Jabbari,
  • Raimund Kirner,
  • Jun Yu

摘要

This paper presents a memetic approach of Competitive Swarm Optimizer (CSO) named Self-Adaptive Competitive Swarm Optimizer (SACSO). We focus on two key parameters of CSO: scaling factor and population size and introduce a parameter sorting scheme and a linear population reduction strategy to enhance the performance of SACSO. Specifically, the parameter sorting scheme assigns smaller scaling factors to better-performing particle individuals to encourage exploitative searchability, while larger scaling factors are allocated to poor-performing particle individuals to promote exploration in unknown search areas. The linear population reduction strategy adapts to different optimization phases, emphasizing exploration in the early phase and enhancing exploitation in the later phase. To investigate the performance of SACSO, we conducted comprehensive numerical experiments on CEC2017, CEC2022, and seven classic engineering problems. The experimental results and statistical analysis demonstrate the competitiveness of SACSO against six state-of-the-art optimizers and four advanced CSO variants. Ablation studies further validate the contribution of the two introduced strategies. Finally, we applied SACSO to real-world Human-Powered Aircraft (HPA) design problems, and the results highlight the potential of SACSO for solving various real-world optimization challenges. The source code of SACSO is available at https://github.com/RuiZhong961230/SACSO.