Scholarlink

<p>An engineering-oriented multi-objective walrus optimizer (MOWO) is proposed for solving constrained multi-objective optimization problems in offshore structural design. Built upon the walrus optimizer, MOWO integrates fast non-dominated sorting, crowding distance-based diversity preservation, and an elitist selection mechanism to perform efficient Pareto-based multi-objective search. The proposed algorithm is comprehensively validated on 32 case studies, covering classical unconstrained benchmark suites (ZDT, DTLZ, and WFG), constrained benchmarks (CEC2009), and 10 real-world engineering design problems. Comprehensive comparative analyses are conducted against 18 well-known multi-objective optimization algorithms. Furthermore, an automated optimization workflow is developed by coupling MOWO with SACS via Python-based secondary development and is applied to the multi-objective optimization of a deep-water jacket platform under extreme environmental load combinations. The obtained Rank 1 Pareto set yields a large number of feasible design alternatives, revealing that the primary trade-off is dominated by structural weight versus member utilization, while displacement and pile utilization exhibit limited variation within the feasible design region. The source code of MOWO is publicly available at <a href="https://github.com/Yuanqyzs/MOWO">https://github.com/Yuanqyzs/MOWO</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

MOWO: A novel multi-objective walrus optimizer for constrained engineering problems

  • 期刊论文

  • 发表日期：

    2026年6月20日

查看全文
  • Muxuan Han,
  • Zunfeng Du,
  • Haiming Zhu,
  • Yujie Yuan,
  • Haitao Zhu,
  • Yancang Li,
  • Qiuyu Yuan

摘要

An engineering-oriented multi-objective walrus optimizer (MOWO) is proposed for solving constrained multi-objective optimization problems in offshore structural design. Built upon the walrus optimizer, MOWO integrates fast non-dominated sorting, crowding distance-based diversity preservation, and an elitist selection mechanism to perform efficient Pareto-based multi-objective search. The proposed algorithm is comprehensively validated on 32 case studies, covering classical unconstrained benchmark suites (ZDT, DTLZ, and WFG), constrained benchmarks (CEC2009), and 10 real-world engineering design problems. Comprehensive comparative analyses are conducted against 18 well-known multi-objective optimization algorithms. Furthermore, an automated optimization workflow is developed by coupling MOWO with SACS via Python-based secondary development and is applied to the multi-objective optimization of a deep-water jacket platform under extreme environmental load combinations. The obtained Rank 1 Pareto set yields a large number of feasible design alternatives, revealing that the primary trade-off is dominated by structural weight versus member utilization, while displacement and pile utilization exhibit limited variation within the feasible design region. The source code of MOWO is publicly available at https://github.com/Yuanqyzs/MOWO.