Scholarlink

<p>The Traveling Salesman Problem (TSP) is a classic combinatorial optimization problem in computer science and operations research. It is a well-known NP-hard problem, and its model is widely present in logistics, transportation, and printed circuit board design. However, traditional methods for solving the TSP often struggle with local optima due to the lack of global perspective and reliance on local decisions. To address this issue, we propose an Adaptive Annealing Dynamic Q-learning Based on Subroute Optimization algorithm for solving the TSP. This algorithm introduces an annealing formula and 2-opt optimization to the traditional Q-learning approach, enabling better escape from local optima. Experimental results show that, compared to traditional Q-learning, the proposed algorithm reduces the path length by 37.93% and also demonstrates improvements of 6.05 to 15.87% over several 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

Adaptive annealing dynamic Q-learning based on subroute optimization for solving the traveling salesman problem

  • 期刊论文

  • 发表日期：

    2025年7月31日

查看全文
  • Liming Sun,
  • Ming Zhu,
  • Jing Li,
  • Haotian Zhang

摘要

The Traveling Salesman Problem (TSP) is a classic combinatorial optimization problem in computer science and operations research. It is a well-known NP-hard problem, and its model is widely present in logistics, transportation, and printed circuit board design. However, traditional methods for solving the TSP often struggle with local optima due to the lack of global perspective and reliance on local decisions. To address this issue, we propose an Adaptive Annealing Dynamic Q-learning Based on Subroute Optimization algorithm for solving the TSP. This algorithm introduces an annealing formula and 2-opt optimization to the traditional Q-learning approach, enabling better escape from local optima. Experimental results show that, compared to traditional Q-learning, the proposed algorithm reduces the path length by 37.93% and also demonstrates improvements of 6.05 to 15.87% over several state-of-the-art algorithms.