Scholarlink

<p>This work introduces two enhanced metaheuristic algorithms, the Modified Opposition-Based Jellyfish Search (MOJS) and the Opposition-Based Non-Boundary Jellyfish Search (ONBJS), both integrating Opposition-Based Learning (OBL) to improve population diversity and balance exploration and exploitation. The algorithms were evaluated on 14 large-scale optimization problems from CEC’2013 and 10 single-objective bound-constrained problems from CEC’2022. The results indicate that ONBJS achieved the most robust and consistent performance on the CEC’2013 benchmark, attaining superior minimum fitness values across multiple functions. Conversely, MOJS excelled in the CEC’2022 benchmark, demonstrating higher stability and better average performance, particularly in multimodal and composite functions. Execution time analysis also revealed a reasonable trade-off between computational cost and solution quality, with MOJS and ONBJS requiring more time due to their enhanced mechanisms, yet consistently delivering superior results. Statistical analyses confirmed the significant advantage of both approaches over classical variants, reinforcing the effectiveness of opposition-based strategies in solving complex optimization problems.</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

Robust Opposition-Based Jellyfish Search Algorithms for Large-Scale and Bound-Constrained Optimization

  • 期刊论文

  • 发表日期：

    2025年9月15日

查看全文
  • Douglas Rodrigues,
  • Luiz Fernando M. O. Sementille,
  • João Paulo Papa

摘要

This work introduces two enhanced metaheuristic algorithms, the Modified Opposition-Based Jellyfish Search (MOJS) and the Opposition-Based Non-Boundary Jellyfish Search (ONBJS), both integrating Opposition-Based Learning (OBL) to improve population diversity and balance exploration and exploitation. The algorithms were evaluated on 14 large-scale optimization problems from CEC’2013 and 10 single-objective bound-constrained problems from CEC’2022. The results indicate that ONBJS achieved the most robust and consistent performance on the CEC’2013 benchmark, attaining superior minimum fitness values across multiple functions. Conversely, MOJS excelled in the CEC’2022 benchmark, demonstrating higher stability and better average performance, particularly in multimodal and composite functions. Execution time analysis also revealed a reasonable trade-off between computational cost and solution quality, with MOJS and ONBJS requiring more time due to their enhanced mechanisms, yet consistently delivering superior results. Statistical analyses confirmed the significant advantage of both approaches over classical variants, reinforcing the effectiveness of opposition-based strategies in solving complex optimization problems.