航空学报 > 2026, Vol. 47 Issue (2): 332097-332097   doi: 10.7527/S1000-6893.2025.32097

结合DM-KM分组的TAS机制增量式调度方法

景世龙1, 施睿2, 周璇3, 闫嘉伟4, 何锋1()   

  1. 1.北京航空航天大学 电子信息工程学院,北京 100083
    2.中国运载火箭技术研究院 空间物理重点实验室,北京 100076
    3.中央民族大学 信息工程学院,北京 100081
    4.中国运载火箭技术研究院 研究发展中心,北京 100076
  • 收稿日期:2025-04-09 修回日期:2025-05-06 接受日期:2025-06-16 出版日期:2025-07-28 发布日期:2025-07-25
  • 通讯作者: 何锋 E-mail:fenghe@buaa.edu.cn
  • 基金资助:
    国家自然科学基金(U2333213);国家自然科学基金(62301014);国家自然科学基金(62071023)

Incremental TAS scheduling method with DM-KM grouping

Shilong JING1, Rui SHI2, Xuan ZHOU3, Jiawei YAN4, Feng HE1()   

  1. 1.School of Electronic and Information Engineer,Beihang University,Beijing 100083,China
    2.Key Laboratory of Science and Technology on Space Physics,China Academy of Launch Vehicle Technology,Beijing 100076,China
    3.School of Information Engineering,Minzu University of China,Beijing 100081,China
    4.Research & Development Center,China Academy of Lunch Vehicle Technology,Beijing 100076,China
  • Received:2025-04-09 Revised:2025-05-06 Accepted:2025-06-16 Online:2025-07-28 Published:2025-07-25
  • Contact: Feng HE E-mail:fenghe@buaa.edu.cn
  • Supported by:
    National Natural Foundation of China(U2333213)

摘要:

针对航天器大规模时间敏感组网应用中面临的时间感知调度(TAS)调度求解规模较低、速度较慢的问题,提出了一种基于距离矩阵和K-means聚类分组的DM-KM流量分组算法,并完成与之结合的增量式TAS调度方法设计。首先,构建流量网络模型,使用基于熵权法的加权综合距离矩阵表示流量之间的相关性。在该模型的基础上,设计并实现了结合DM-KM流量分组的增量式调度算法。所提出的分组方法具有较大的组内相似性和较低的组间相似性,流量分组能够有效提升增量式调度求解速度。实验结果表明:与现有DoC-KM和CILP-KM分组算法相比,在1 000条流量调度场景下,DM-KM算法在维持较高求解速度的基础上,拥有较好的可调度性。相较于其他调度算法,求解规模提升最大可达32.36%,为TSN网络在航天器大规模组网提供了分组增量式的调度解决方案。

关键词: 时间敏感网络(TSN), 时间感知调度(TAS), 流量分组, 增量式求解框架, 箭载网络

Abstract:

In large-scale, time-sensitive networking applications for spacecraft, the Time-Aware Scheduling (TAS) scheduling often faces challenges such as relatively low solving scale and slow speed. This paper proposes a DM-KM traffic grouping algorithm based on a distance matrix and K-means clustering, and integrated with it, designs an incremental TAS scheduling method. First, a traffic network model is constructed, using a weighted comprehensive distance matrix based on the entropy weight method to represent the correlations between traffic flows. Then, an incremental scheduling algorithm combined with DM-KM traffic grouping is designed and implemented. The proposed grouping method achieves high intra-group similarity and low inter-group similarity, which effectively improves the solving speed of the incremental scheduling. Experimental results show that compared with the existing DoC-KM and CILP-KM grouping algorithms, the DM-KM algorithm achieves better schedulability while maintaining a high solving speed in a 1000-traffic scheduling scenario. Compared with other scheduling algorithms, the solving scale can be improved by up to 32.36%, providing a grouping and incremental scheduling solution for Time-Sensitive Networking (TSN) in large-scale spacecraft networks.

Key words: Time-Sensitive Network (TSN), Time-Aware Shaper (TAS), flow grouping, incremental solving framework, on-board network

中图分类号: