Scholarlink

<p>This study offers a multi-strategy improved snake optimizer (MISO) and applies it to library robot path&#xa0;planning in order to solve the limitations of the conventional snake optimizer, including slow convergence, low solution accuracy, and susceptibility to local optima in complex optimization problems. Firstly, the tent mapping has been used to initialize the population and enhance randomness and diversity, thereby enhancing global optimization capabilities. Secondly, a dynamic inertia weight factor is combined with Lévy flight to optimize the search performance in the exploration phase and balance global search and local exploitation. Thirdly, the centroid opposition-based learning has been used to update individual positions and maintain diversity. The performance of MISO was evaluated using the CEC 2005 benchmark test function set. Results show that MISO significantly outperforms ten comparative algorithms in most cases on the 23 test functions. Ablation experiments show the effectiveness of the three strategies. In the application of library robot path planning, three complex environments were designed to test MISO’s practicality. The results show that MISO’s average path length was reduced significantly compared to comparative algorithms, with a target search accuracy rate of over 97%, significantly superior to comparative algorithms. The paper demonstrates that MISO provides a novel technical solution for efficient and safe path planning of robots in smart libraries, and its multi-strategy fusion concept can offer references for other complex optimization problems.</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 improved snake optimizer for library robot path planning problems

  • 期刊论文

  • 发表日期：

    2025年7月14日

查看全文
  • Yige Xue

摘要

This study offers a multi-strategy improved snake optimizer (MISO) and applies it to library robot path planning in order to solve the limitations of the conventional snake optimizer, including slow convergence, low solution accuracy, and susceptibility to local optima in complex optimization problems. Firstly, the tent mapping has been used to initialize the population and enhance randomness and diversity, thereby enhancing global optimization capabilities. Secondly, a dynamic inertia weight factor is combined with Lévy flight to optimize the search performance in the exploration phase and balance global search and local exploitation. Thirdly, the centroid opposition-based learning has been used to update individual positions and maintain diversity. The performance of MISO was evaluated using the CEC 2005 benchmark test function set. Results show that MISO significantly outperforms ten comparative algorithms in most cases on the 23 test functions. Ablation experiments show the effectiveness of the three strategies. In the application of library robot path planning, three complex environments were designed to test MISO’s practicality. The results show that MISO’s average path length was reduced significantly compared to comparative algorithms, with a target search accuracy rate of over 97%, significantly superior to comparative algorithms. The paper demonstrates that MISO provides a novel technical solution for efficient and safe path planning of robots in smart libraries, and its multi-strategy fusion concept can offer references for other complex optimization problems.