Scholarlink

Multi-objective particle swarm optimization (MOPSO) emerged as a robust approach to solve multi-objective optimization problems (MOPs). A comprehensive review and empirical evaluation of eight state-of-the-art MOPSO algorithms is conducted. These decomposition-based, dominance-based, and criterion-based approaches are analysed using benchmark functions from the Zitzler Deb Thiele (ZDT) and the Walking Fish Group (WFG) suites. Based on rigorous statistical tests, the competitive mechanism-based MOPSO and the MOPSO with multiple search strategies demonstrate superior performance across the different MOPs, while decomposition-based MOPSO and multi-objective cooperative PSO exhibit scalability limitations. The results contribute to a better understanding of MOPSO approaches, offering recommendations for selecting the most suitable algorithm for MOPs.

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

An Empirical Analysis of Particle Swarm Optimisation Approaches for Multi-objective Optimisation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Lauren Trinder-Smith,
  • Andries P. Engelbrecht

摘要

Multi-objective particle swarm optimization (MOPSO) emerged as a robust approach to solve multi-objective optimization problems (MOPs). A comprehensive review and empirical evaluation of eight state-of-the-art MOPSO algorithms is conducted. These decomposition-based, dominance-based, and criterion-based approaches are analysed using benchmark functions from the Zitzler Deb Thiele (ZDT) and the Walking Fish Group (WFG) suites. Based on rigorous statistical tests, the competitive mechanism-based MOPSO and the MOPSO with multiple search strategies demonstrate superior performance across the different MOPs, while decomposition-based MOPSO and multi-objective cooperative PSO exhibit scalability limitations. The results contribute to a better understanding of MOPSO approaches, offering recommendations for selecting the most suitable algorithm for MOPs.