航空学报 > 2025, Vol. 46 Issue (13): 531274-531274   doi: 10.7527/S1000-6893.2024.31274

差异化保障需求驱动的舰载机多机协同决策方法

陈伟1, 李璐璐1, 陈董1,2,3, 张少辉1,4, 李亚飞1,2,3(), 王可1,2,3, 靳远远1,2,3, 徐明亮1,2,3   

  1. 1.郑州大学 计算机与人工智能学院,郑州 450001
    2.智能集群系统教育部工程研究中心,郑州 450001
    3.国家超级计算郑州中心,郑州 450001
    4.周口师范学院 人工智能学院,周口 466001
  • 收稿日期:2024-09-27 修回日期:2024-10-24 接受日期:2024-12-06 出版日期:2024-12-24 发布日期:2024-12-23
  • 通讯作者: 李亚飞 E-mail:ieyfli@zzu.edu.cn
  • 基金资助:
    国家自然科学基金(62372416);国家自然科学基金(62302460);国家自然科学基金(61972362);国家自然科学基金(62036010);国家自然科学基金(62325602);河南省自然科学基金(242300421215);中国博士后科学基金(2022TQ0297);中国博士后科学基金(2020M682348)

Multi-aircraft cooperative decision-making methods driven by differentiated support demands for carrier-based aircraft

Wei CHEN1, Lulu LI1, Dong CHEN1,2,3, Shaohui ZHANG1,4, Yafei LI1,2,3(), Ke WANG1,2,3, Yuanyuan JIN1,2,3, Mingliang XU1,2,3   

  1. 1.School of Computer Science and Artificial Intelligence,Zhengzhou University,Zhengzhou 450001,China
    2.Intelligent Cluster System Engineering Research Center of the Ministry of Education,Zhengzhou 450001,China
    3.National Supercomputing Center in Zhengzhou,Zhengzhou 450001,China
    4.School of Artificial Intelligence,Zhoukou Normal University,Zhoukou 466001,China
  • Received:2024-09-27 Revised:2024-10-24 Accepted:2024-12-06 Online:2024-12-24 Published:2024-12-23
  • Contact: Yafei LI E-mail:ieyfli@zzu.edu.cn
  • Supported by:
    National Natural Science Foundation of China(62372416);Natural Science Foundation of Henan(242300421215);China Postdoctoral Science Foundation(2022TQ0297)

摘要:

多机种舰载机编队作战是现代航母作战的核心模式,保障能力是决定现代海战胜负的关键。然而,实战中多机种舰载机保障过程面临着多重挑战,包括舰载机保障流程的差异性、保障资源的多元化约束以及作业工序的复杂性。为应对这些挑战,提出了一种新的依赖感知的任务决策调度模型DATSDM。该模型结合图神经网络,深入剖析保障流程的网络结构关系,实现了跨规模、多机种集群保障资源的高效调度。同时,DATSDM还融合了Transformer模型,能够并行处理舰载机保障资源调度任务,大幅缩短多机协同保障时间。实验结果显示,DATSDM在大规模舰载机资源配置场景中表现卓越,相比同类算法,能将多机种舰载机编队保障时间降低约13.36%,提升了多机种舰载机协同保障效率与作战效能。

关键词: 舰载机, 深度强化学习, 图神经网络, 调度优化, 资源分配

Abstract:

In modern naval warfare, the effectiveness of carrier-borne aircraft operations is crucial, and multi-type carrier-borne aircraft formations have become the fundamental operational paradigm for aircraft carriers. However, collaborative decision making for carrier-borne aircraft support faces significant challenges due to differentiated support processes, deck space limitations, and complex maintenance procedures. To address these scheduling challenges, we propose a novel Dependency-Aware Task Scheduling Decision Module (DATSDM). By leveraging graph neural networks, DATSDM delves into the intricate network structure of support processes, enabling efficient and precise scheduling of support resources across varying scales and multi-type carrier-borne aircraft clusters. Furthermore, DATSDM incorporates the strengths of transformer, harnessing its attention mechanism to parallelly analyze and process carrier-borne aircraft support information, thereby significantly reducing support times. Extensive experiments demonstrate the remarkable superiority of DATSDM over its peers. In various resource allocation scenarios for multiple carrier-borne aircraft types, DATSDM reduces support time by 13.36%. This improvement significantly enhances the overall support efficiency and combat readiness of multi-type carrier-borne aircraft.

Key words: carrier-based aircraft, deep reinforcement learning, graph neural networks, scheduling optimization, resource allocation

中图分类号: