基于蒙特卡洛树搜索的舰载机保障作业调度方法

  • 彭健 ,
  • 李亚飞 ,
  • 吴庆顺 ,
  • 朱广磊 ,
  • 贺硕 ,
  • 靳远远 ,
  • 徐明亮
展开
  • 1. 郑州大学计算机与人工智能学院
    2. 郑州大学
    3. 郑州大学 计算机与人工智能学院

收稿日期: 2025-06-19

  修回日期: 2025-09-19

  网络出版日期: 2025-09-24

基金资助

国家自然科学基金杰青项目;国家自然科学基金面上项目;国家自然科学基金重点项目;国家自然科学基金面上项目;河南省自然科学基金重点项目;国家自然科学青年项目;国家自然科学青年项目;中国博士后科学基金

Scheduling method for carrier-based aircraft support operations based on Monte Carlo tree search

  • PENG Jian ,
  • LI Ya-Fei ,
  • WU Qing-Shun ,
  • ZHU Guang-Lei ,
  • HE Shuo ,
  • JIN Yuan-Yuan ,
  • XU Ming-Liang
Expand

Received date: 2025-06-19

  Revised date: 2025-09-19

  Online published: 2025-09-24

摘要

制定高效的航母舰载机甲板保障作业调度计划是提升舰载机出动效能的关键。为了提升机群保障能力,对航母舰载机甲板保障作业调度算法进行了研究。首先,针对舰载机甲板保障作业调度问题的任务需求以及各种约束条件,建立了以最小化机群保障完工时间为优化目标的约束满足模型,并且对该调度问题进行计算复杂性分析,确定其属于NP-hard问题。然后,针对该问题提出了基于双向对齐技术的调度方案生成方法,并设计了一种融合蒙特卡洛树搜索的启发式算法来优化作业序列。该算法借鉴了蒙特卡洛树搜索的探索与利用平衡机制,结合启发式模拟策略和调度生成方法来评估搜索路径,利用蒙特卡洛树记录评估结果以引导后续搜索方向。最后,为了验证所提出算法的性能,使用随机活动网络生成器构造测试案例并开展仿真实验。实验结果表明,相较于当前先进算法,所提算法的求解质量与求解效率均有提升。

本文引用格式

彭健 , 李亚飞 , 吴庆顺 , 朱广磊 , 贺硕 , 靳远远 , 徐明亮 . 基于蒙特卡洛树搜索的舰载机保障作业调度方法[J]. 航空学报, 0 : 1 -0 . DOI: 10.7527/S1000-6893.2025.32444

Abstract

The key to improving the takeoff efficiency of carrier aircraft is to develop an efficient deck support operation scheduling plan for carrier aircraft. In order to improve the aircraft group support ability, the aircraft carrier deck support operation scheduling algorithm is studied. Firstly, according to the task requirements and various constraints of the carrier-based aircraft deck support scheduling problem, a constraint satisfaction model is established to minimize the completion time of cluster support, and the computational complexity analysis of the scheduling problem shows that it is an NP-hard problem. The algorithm draws on the exploration and exploitation balance mechanism of Monte Carlo tree search, combines heuristic simulation strategy and scheduling scheme generation method to evaluate the search path, and uses the Monte Carlo tree to record the evaluation results to guide the subsequent search direction. Finally, to verify the performance of the proposed algorithm, a random activity network generator is used to construct test cases and conduct simulation experiments. The experimental results show that compared with the current advanced algorithms, the proposed algorithm improves the solution quality and efficiency.
文章导航

/