Scholarlink

<p>Many metaheuristic algorithms rely too heavily on metaphorical sources of inspiration, resulting in unnecessary complexity and redundant operations. This contradicts the core goal of algorithm design: simplicity and efficiency. This paper attempts to break free from the constraints of inspiration and presents a simple yet efficient metaheuristic algorithm: Delta Plus. Specifically, we propose a Delta operation to drive the Delta Plus for global optimization. The key features of the Delta operation are: (1) it does not reliance on the best solution to guide the search, and (2) it considers both the current and previous iteration’s solution changes. We evaluate the Delta Plus and seven recent metaheuristics on the CEC2017 benchmark, as well as the Delta Plus and five state-of-the-art algorithms on the CEC2022 benchmarks. Finally, the Delta Plus and seven recent metaheuristics are applied to solve five mechanical engineering problems. The results indicate that the proposed Delta Plus can be considered a simple yet efficient metaheuristic algorithm. The code for the proposed Delta Plus is available at <a href="https://ww2.mathworks.cn/matlabcentral/fileexchange/177839-delta-plus">https://ww2.mathworks.cn/matlabcentral/fileexchange/177839-delta-plus</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

Freedom from inspiration! Achieving efficient metaheuristic optimization with delta plus

  • 期刊论文

  • 发表日期：

    2025年7月31日

查看全文
  • Yuansheng Gao,
  • Jinpeng Wang,
  • Lang Qin,
  • Jiahui Zhang,
  • Yulin Wang

摘要

Many metaheuristic algorithms rely too heavily on metaphorical sources of inspiration, resulting in unnecessary complexity and redundant operations. This contradicts the core goal of algorithm design: simplicity and efficiency. This paper attempts to break free from the constraints of inspiration and presents a simple yet efficient metaheuristic algorithm: Delta Plus. Specifically, we propose a Delta operation to drive the Delta Plus for global optimization. The key features of the Delta operation are: (1) it does not reliance on the best solution to guide the search, and (2) it considers both the current and previous iteration’s solution changes. We evaluate the Delta Plus and seven recent metaheuristics on the CEC2017 benchmark, as well as the Delta Plus and five state-of-the-art algorithms on the CEC2022 benchmarks. Finally, the Delta Plus and seven recent metaheuristics are applied to solve five mechanical engineering problems. The results indicate that the proposed Delta Plus can be considered a simple yet efficient metaheuristic algorithm. The code for the proposed Delta Plus is available at https://ww2.mathworks.cn/matlabcentral/fileexchange/177839-delta-plus.