Page 162 - 《软件学报》2026年第5期

廖家俊 等: TB-Match: 融合条件扩散模型与近端策略优化的弹性时间约束运单分配方法 2041

可扩展性, 对于大规模的运单分配任务消耗了较长的时间. TB-Match 稳定的 MR 表明其学习到的策略泛化良好,
并且该框架可以处理分配规模的大幅增加, 而不会产生过高的计算开销或匹配质量的显著损失.

表 8 可扩展性分析

TB-Match MAPT HA
数据集规模 (运单数) (k)
时间 (ms/批次) MR (%) 时间 (ms/批次) MR (%) 时间 (ms/批次) MR (%)
50 (基线) 817 92.3 624 89.1 4 573 81.5
100 1 489 92.1 1 106 88.7 8 581 81.0
200 2 270 91.8 1 958 88.2 15 402 80.2
500 5 096 91.2 4 125 87.5 33 957 78.5

5 总 结

本文提出的 TB-Match 方法解决了网络货运平台运单分配中的两个关键问题: 刚性时间约束的弹性化建模以
及匹配率与司机满意度的多目标优化. 该方法通过 4 个协同工作的技术模块实现了系统性能的提升: 基于条件扩
散模型的概率化时间约束评估、融合多维特征的司机偏好表征、强化学习驱动的动态权重调节以及近端策略优
化的长期决策生成. 在日均 30 000 吨货运量的实际运营场景中, TB-Match 方法实现了 17.66% 的匹配率提升, 验
证了其在实际应用中的有效性. 未来的研究工作将从多个方向展开进一步拓展: 首先, 通过整合天气、路况等动态
外部因素增强模型的环境适应性; 其次, 开发适用于冷启动场景的迁移学习策略, 优化新司机和新区域的分配效率;
最后, 探索可持续的智能分配范式, 推动物流平台向更高效、更智能的方向发展. 这些研究方向将进一步强化智能
物流系统的鲁棒性和普适性.

References
[1] Ge Y, Xiong H, Tuzhilin A, Xiao KL, Gruteser M, Pazzani M. An energy-efficient mobile recommender system. In: Proc. of the 16th
ACM SIGKDD Int’l Conf. on Knowledge Discovery and Data Mining. Washington: ACM, 2010. 899–908. [doi: 10.1145/1835804.
1835918]
[2] Stiglic M, Agatz N, Savelsbergh M, Gradisar M. The benefits of meeting points in ride-sharing systems. Transportation Research Part B:
Methodological, 2015, 82: 36–53. [doi: 10.1016/j.trb.2015.07.025]
[3] Zhao Y, Xia JF, Liu GF, Su H, Lian DF, Shang S, Zheng K. Preference-aware task assignment in spatial crowdsourcing. In: Proc. of the
33rd AAAI Conf. on Artificial Intelligence. Honolulu: AAAI, 2019. 2629–2636. [doi: 10.1609/aaai.v33i01.33012629]
[4] Kuhn HW. The Hungarian method for the assignment problem. Naval Research Logistics Quarterly, 1955, 2(1–2): 83–97. [doi: 10.1002/
nav.3800020109]
[5] Jonker R, Volgenant A. A shortest augmenting path algorithm for dense and sparse linear assignment problems. Computing, 1987, 38(4):
325–340. [doi: 10.1007/BF02278710]
[6] Burkard RE, Dell’Amico M, Martello S. Assignment Problems. Philadelphia: SIAM, 2009.
[7] Bertsekas DP. The auction algorithm: A distributed relaxation method for the assignment problem. Annals of Operations Research, 1988,
14(1): 105–123. [doi: 10.1007/BF02186476]
[8] Pisinger D, Ropke S. A general heuristic for vehicle routing problems. Computers & Operations Research, 2007, 34(8): 2403–2435. [doi:
10.1016/j.cor.2005.09.012]
[9] Lee S, Boomsma TK. An approximate dynamic programming algorithm for short-term electric vehicle fleet operation under uncertainty.
Applied Energy, 2022, 325: 119793. [doi: 10.1016/j.apenergy.2022.119793]
[10] Yan CW, Zhu HL, Korolko N, Woodard D. Dynamic pricing and matching in ride-hailing platforms. Naval Research Logistics (NRL),
2020, 67(8): 705–724. [doi: 10.1002/nav.21872]
[11] Zhu TC, Qiu Y, Zhou HY, Li JX. Decoding global preferences: Temporal and cooperative dependency modeling in multi-agent
preference-based reinforcement learning. In: Proc. of the 38th AAAI Conf. on Artificial Intelligence. Vancouver: AAAI, 2024.
17202–17210. [doi: 10.1609/aaai.v38i15.29666]
[12] Zhen XH, Long J, Cai ZP. A survey of order dispatch policy based on online ride-hailing services. Computer Engineering and Science,
2020, 42(7): 1267–1275 (in Chinese with English abstract). [doi: 10.3969/j.issn.1007-130X.2020.07.016]