DeepFreight：融合深度强化学习与混合整数规划的多中转卡车货运配送
机器学习
2023-05-26 v2 多智能体系统
摘要
随着货运需求与运输成本的快速上升，对车队进行智能控制以实现高效且注重成本的解决方案成为一个重要问题。本文提出 DeepFreight，一种基于无模型深度强化学习的多中转货运配送算法，其包含两个紧密协作的组件：卡车调度与包裹匹配。具体而言，利用称为 QMIX 的深度多智能体强化学习框架学习调度策略，借此可为车队针对配送请求获得多步联合车辆调度决策。随后执行高效的多中转匹配算法将配送请求分配给卡车。此外，DeepFreight 还集成了混合整数线性规划（Mixed-Integer Linear Programming）优化器以进一步优化。评估结果表明，所提系统具有高度可扩展性，并在保持较低配送时间与燃料消耗的同时确保 100% 配送成功率。代码见 https://github.com/LucasCJYSDL/DeepFreight。
引用
@article{arxiv.2103.03450,
title = {DeepFreight: Integrating Deep Reinforcement Learning and Mixed Integer Programming for Multi-transfer Truck Freight Delivery},
author = {Jiayu Chen and Abhishek K. Umrawal and Tian Lan and Vaneet Aggarwal},
journal= {arXiv preprint arXiv:2103.03450},
year = {2023}
}
备注
Citing the ICAPS version is preferred: Chen, Jiayu, Abhishek K. Umrawal, Tian Lan, and Vaneet Aggarwal. "DeepFreight: A Model-free Deep-reinforcement-learning-based Algorithm for Multi-transfer Freight Delivery." In Proceedings of the International Conference on Automated Planning and Scheduling, vol. 31, pp. 510-518. 2021