Scholarlink

<p>To address the rapid loss of diversity and premature convergence of the Differential Evolution (DE) algorithm in high-dimensional optimization, this paper proposes an improved algorithm, AL-EnhancedDE. The method integrates a dimensional chaotic rebirth strategy and an adaptive step-size search mechanism to enhance robustness in solving complex problems. By incorporating a dynamic disruption mechanism, three repair strategies, and an adaptive chaotic Lévy flight mechanism, the algorithm can effectively approach the global optimum while maintaining stability. The proposed AL-EnhancedDE was evaluated on high-dimensional settings (50 and 100 dimensions) of the CEC2017 benchmark suite. It was compared with classical heuristic algorithms (DE, PSO, EAO, WAA, AOO, DOA, SFOA) and advanced DE variants (JADE, SADE, CoDE, HSDE, LSHADE, ADE, jDE, AJSO, NFDDE, LSHADE-SPACMA, MIDE, QSMODE, MODEA). Experimental results indicate that the AL-EnhancedDE algorithm outperforms the other 20 comparison algorithms in overall performance.</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

An enhanced differential evolution algorithm with dimensional chaos rebirth and adaptive step-size strategy for high dimensional optimization

  • 期刊论文

  • 发表日期：

    2026年2月12日

查看全文
  • Lishu Qin,
  • Xinyuan Lu,
  • Yu Gao

摘要

To address the rapid loss of diversity and premature convergence of the Differential Evolution (DE) algorithm in high-dimensional optimization, this paper proposes an improved algorithm, AL-EnhancedDE. The method integrates a dimensional chaotic rebirth strategy and an adaptive step-size search mechanism to enhance robustness in solving complex problems. By incorporating a dynamic disruption mechanism, three repair strategies, and an adaptive chaotic Lévy flight mechanism, the algorithm can effectively approach the global optimum while maintaining stability. The proposed AL-EnhancedDE was evaluated on high-dimensional settings (50 and 100 dimensions) of the CEC2017 benchmark suite. It was compared with classical heuristic algorithms (DE, PSO, EAO, WAA, AOO, DOA, SFOA) and advanced DE variants (JADE, SADE, CoDE, HSDE, LSHADE, ADE, jDE, AJSO, NFDDE, LSHADE-SPACMA, MIDE, QSMODE, MODEA). Experimental results indicate that the AL-EnhancedDE algorithm outperforms the other 20 comparison algorithms in overall performance.