Scholarlink

<p>The Particle Swarm Optimizer (PSO) algorithm is widely recognized as a leading metaheuristic for solving various optimization problems. However, it often suffers from slow convergence speed and a tendency to fall into local optima. To address these issues, this study proposes a Multi-Strategy Evolution-Enhanced Particle Swarm Optimizer (MEE-PSO), which incorporates an adaptive elimination mechanism and chaotic mapping to dynamically adjust the proportion of elite and ordinary particles at different search stages. This approach enhances population diversity and mitigates homogenization. Additionally, a lateral crossover mechanism further improves the swarm’s exploration capability, while an adaptive mutation strategy enhances local exploitation and prevents premature stagnation. The optimization performance of MEE-PSO is validated through comparative experiments against 32 state-of-the-art algorithms on 151 continuous optimization problems, including 120 CEC2017 functions, 24 CEC2022 functions, and 11 real-world engineering optimization tasks. Furthermore, MEE-PSO is applied to 15 discrete optimization problems, comprising 4 traveling salesman problems, 10 capacitated vehicle routing problems, and a cold chain logistics problem, and its performance is evaluated against 52 outstanding optimization algorithms. Wilcoxon rank-sum tests and Friedman tests indicate that MEE-PSO significantly outperforms competing algorithms on most benchmark sets in both low- and high-dimensional scenarios. In summary, the superior performance of MEE-PSO in solving complex high-dimensional continuous and discrete optimization problems demonstrates its applicability across a wide range of real-world optimization tasks. The code of MEE-PSO can be obtained in <a href="https://github.com/guangian/MEE-PSO">https://github.com/guangian/MEE-PSO</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

Multi-strategy evolutionary enhanced particle swarm optimizer with adaptive morphology for engineering example

  • 期刊论文

  • 发表日期：

    2025年9月22日

查看全文
  • Ziyu Guan,
  • Chenhua Liu,
  • Tao Wang,
  • Haomin Cui,
  • Changjiang Ren,
  • Changhai Qin

摘要

The Particle Swarm Optimizer (PSO) algorithm is widely recognized as a leading metaheuristic for solving various optimization problems. However, it often suffers from slow convergence speed and a tendency to fall into local optima. To address these issues, this study proposes a Multi-Strategy Evolution-Enhanced Particle Swarm Optimizer (MEE-PSO), which incorporates an adaptive elimination mechanism and chaotic mapping to dynamically adjust the proportion of elite and ordinary particles at different search stages. This approach enhances population diversity and mitigates homogenization. Additionally, a lateral crossover mechanism further improves the swarm’s exploration capability, while an adaptive mutation strategy enhances local exploitation and prevents premature stagnation. The optimization performance of MEE-PSO is validated through comparative experiments against 32 state-of-the-art algorithms on 151 continuous optimization problems, including 120 CEC2017 functions, 24 CEC2022 functions, and 11 real-world engineering optimization tasks. Furthermore, MEE-PSO is applied to 15 discrete optimization problems, comprising 4 traveling salesman problems, 10 capacitated vehicle routing problems, and a cold chain logistics problem, and its performance is evaluated against 52 outstanding optimization algorithms. Wilcoxon rank-sum tests and Friedman tests indicate that MEE-PSO significantly outperforms competing algorithms on most benchmark sets in both low- and high-dimensional scenarios. In summary, the superior performance of MEE-PSO in solving complex high-dimensional continuous and discrete optimization problems demonstrates its applicability across a wide range of real-world optimization tasks. The code of MEE-PSO can be obtained in https://github.com/guangian/MEE-PSO.