Scholarlink

<p>Global optimization solutions depend on maintaining an equilibrium between exploration and exploitation to efficiently discover optimal results within the search space.&#xa0;Various fields, including engineering and mechanical system design, face these problems because constraints dictate feasible regions. The complexity inherent in optimization problems arises from a combination of interacting factors, including multiple local optima, irregular search landscapes, and symmetrical patterns that influence scaling sensitivity. In addition, interdependent design variables and intricate basin structures often create misleading search cues, further complicating the optimization process. To enhance performance and mitigate these challenges, the Special Relativity Search (SRS) algorithm is integrated with Differential Evolution (DE). This hybridization leads to faster convergence and improved solution quality. The SRS algorithm draws its conceptual foundation from Einstein’s theory of special relativity by emulating the motion of charged particles within a magnetic field. Within this framework, the search space is modeled as a magnetic field, and the candidate solutions behave as charged particles subjected solely to the magnetic component of the Lorentz force. The incorporation of dynamic guidance significantly enhances search efficiency, rendering the method suitable for complex engineering applications such as structural design and control optimization. The proposed approach was evaluated on two structurally complex optimization problems involving both continuous and discrete design variables. Performance comparisons among algorithms were conducted through statistical analyses. The results clearly demonstrate that the proposed hybrid algorithm consistently outperforms existing methods in terms of accuracy, stability, and convergence behavior.</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

Accelerating the convergence of the special relativity search algorithm: a novel hybrid approach for global optimization

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月24日

查看全文
  • Vahid Goodarzimehr,
  • Afshar Seifiasl,
  • Nima Khodadadi

摘要

Global optimization solutions depend on maintaining an equilibrium between exploration and exploitation to efficiently discover optimal results within the search space. Various fields, including engineering and mechanical system design, face these problems because constraints dictate feasible regions. The complexity inherent in optimization problems arises from a combination of interacting factors, including multiple local optima, irregular search landscapes, and symmetrical patterns that influence scaling sensitivity. In addition, interdependent design variables and intricate basin structures often create misleading search cues, further complicating the optimization process. To enhance performance and mitigate these challenges, the Special Relativity Search (SRS) algorithm is integrated with Differential Evolution (DE). This hybridization leads to faster convergence and improved solution quality. The SRS algorithm draws its conceptual foundation from Einstein’s theory of special relativity by emulating the motion of charged particles within a magnetic field. Within this framework, the search space is modeled as a magnetic field, and the candidate solutions behave as charged particles subjected solely to the magnetic component of the Lorentz force. The incorporation of dynamic guidance significantly enhances search efficiency, rendering the method suitable for complex engineering applications such as structural design and control optimization. The proposed approach was evaluated on two structurally complex optimization problems involving both continuous and discrete design variables. Performance comparisons among algorithms were conducted through statistical analyses. The results clearly demonstrate that the proposed hybrid algorithm consistently outperforms existing methods in terms of accuracy, stability, and convergence behavior.