Scholarlink

<p>The dandelion optimizer (DO) is an advanced swarm intelligence algorithm, but still exhibits certain limitations. This paper proposes an enhanced DO named EFDO. Firstly, a hybrid piecewise logistic-circle map is proposed to generate a uniform and high-quality initial population with superior ergodicity and diversity. Secondly, based on elliptic curves, we propose a novel approximation strategy for the first time, and integrate it into the DO, which improves the solution accuracy, and effectively assists the algorithm in avoiding entrapment in local optima. Thirdly, we innovatively incorporate a new similarity metric operator into the original fitness-distance balance method, creating a novel selection strategy named adaptive fitness-distance-similarity balance. This new strategy can effectively explore potential excellent solutions, increase the diversity and prevent premature convergence. EFDO achieves better performance compared against twelve algorithms on the CEC2017 benchmark&#xa0;functions and six engineering problems. Finally, EFDO is applied to parameter estimation of photovoltaic models, underscoring its application capability.</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

Multi-strategy enhanced dandelion optimizer based on elliptic approximation strategy and adaptive fitness-distance-similarity balance for solar photovoltaic parameter estimation

  • 期刊论文

  • 发表日期：

    2025年2月3日

查看全文
  • Tianbao Liu,
  • Zhe Feng

摘要

The dandelion optimizer (DO) is an advanced swarm intelligence algorithm, but still exhibits certain limitations. This paper proposes an enhanced DO named EFDO. Firstly, a hybrid piecewise logistic-circle map is proposed to generate a uniform and high-quality initial population with superior ergodicity and diversity. Secondly, based on elliptic curves, we propose a novel approximation strategy for the first time, and integrate it into the DO, which improves the solution accuracy, and effectively assists the algorithm in avoiding entrapment in local optima. Thirdly, we innovatively incorporate a new similarity metric operator into the original fitness-distance balance method, creating a novel selection strategy named adaptive fitness-distance-similarity balance. This new strategy can effectively explore potential excellent solutions, increase the diversity and prevent premature convergence. EFDO achieves better performance compared against twelve algorithms on the CEC2017 benchmark functions and six engineering problems. Finally, EFDO is applied to parameter estimation of photovoltaic models, underscoring its application capability.