Scholarlink

<p>The Tunicate Swarm Algorithm (TSA), a recent swarm-intelligence optimization approach, has gained significant prominence in various fields. However, its performance on complex multi-modal optimization problems remains suboptimal due to limitations such as low search accuracy, inadequate search capabilities, and a propensity to get stuck in the local optima. These issues arise primarily from the rapid loss of population diversity during the optimization process, resulting in imbalanced exploration and exploitation. To mitigate these issues, this study presents the Augmented Tunicate Swarm Algorithm (ATSA) by embedding two key enhancements: (1) the utilization of a Halton sequence with low-dispersion for population initialization, which enhances the diversity of initial solutions; and (2) the incorporation of a novel Exponential Local Escaping Operator (ELEO) that leverages an exponential mechanism to enhance global search capability, improve convergence behavior, and strengthen the algorithm’s ability to jump out of the local optima. These techniques assist in a good trade-off between diversification and intensification of the algorithm. The proposed ATSA approach is comprehensively evaluated on a set of 23 standard CEC’05, 10 complex CEC’21, and 12 recently developed CEC’22 test functions, benchmarked against eleven state-of-the-art algorithms. The quantitative results reveal that ATSA outperforms competing algorithms on 96% of the CEC’05, 60% of the CEC’21, and 50% of the CEC’22 functions. To further assess its practical applicability, ATSA has also been applied to four real-world constrained engineering optimization problems. Moreover, rigorous statistical analyses, including the Friedman test and pairwise Wilcoxon rank-sum test, confirm ATSA’s competitive advantage, with first-place rankings in 61% of CEC’05, 50% of CEC’21, and 33% of CEC’22 functions, highlighting its robustness across diverse problems. Overall, the experimental findings demonstrate that ATSA surpasses the basic TSA and other competing algorithms in terms of convergence speed, optimization ability, and solution stability. Moreover, ATSA has achieved a general level in algorithm complexity and exhibits competitive performance with other state-of-the-art algorithms.</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

Atsa: a novel augmented low-discrepancy sequence initialized tunicate swarm-based exponential local escaping operator for engineering design applications

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月3日

查看全文
  • Vanisree Chandran,
  • Prabhujit Mohapatra

摘要

The Tunicate Swarm Algorithm (TSA), a recent swarm-intelligence optimization approach, has gained significant prominence in various fields. However, its performance on complex multi-modal optimization problems remains suboptimal due to limitations such as low search accuracy, inadequate search capabilities, and a propensity to get stuck in the local optima. These issues arise primarily from the rapid loss of population diversity during the optimization process, resulting in imbalanced exploration and exploitation. To mitigate these issues, this study presents the Augmented Tunicate Swarm Algorithm (ATSA) by embedding two key enhancements: (1) the utilization of a Halton sequence with low-dispersion for population initialization, which enhances the diversity of initial solutions; and (2) the incorporation of a novel Exponential Local Escaping Operator (ELEO) that leverages an exponential mechanism to enhance global search capability, improve convergence behavior, and strengthen the algorithm’s ability to jump out of the local optima. These techniques assist in a good trade-off between diversification and intensification of the algorithm. The proposed ATSA approach is comprehensively evaluated on a set of 23 standard CEC’05, 10 complex CEC’21, and 12 recently developed CEC’22 test functions, benchmarked against eleven state-of-the-art algorithms. The quantitative results reveal that ATSA outperforms competing algorithms on 96% of the CEC’05, 60% of the CEC’21, and 50% of the CEC’22 functions. To further assess its practical applicability, ATSA has also been applied to four real-world constrained engineering optimization problems. Moreover, rigorous statistical analyses, including the Friedman test and pairwise Wilcoxon rank-sum test, confirm ATSA’s competitive advantage, with first-place rankings in 61% of CEC’05, 50% of CEC’21, and 33% of CEC’22 functions, highlighting its robustness across diverse problems. Overall, the experimental findings demonstrate that ATSA surpasses the basic TSA and other competing algorithms in terms of convergence speed, optimization ability, and solution stability. Moreover, ATSA has achieved a general level in algorithm complexity and exhibits competitive performance with other state-of-the-art algorithms.