| 1,676 | 3 | 1167 |
| 下载次数 | 被引频次 | 阅读次数 |
【背景】随着智慧城市建设和低空经济的快速崛起，城市多机巢无人机配送系统在智能物流与城市低空经济发展中潜力巨大，但是同时面临路径规划、静态禁飞区避障与恒定风场影响等复杂问题，这些因素显著增加了调度优化的难度。【目标】针对多机巢无人机配送中的复杂约束，设计路径配送算法，以实现配送路径最短、能耗最低且避障安全的综合调度优化。【方法】构建多目标混合整数规划模型，综合考虑配送需求、飞行时间窗、静态禁飞区避让与风场影响，引入多智能体图神经网络算法（MAS-GNN），融合图神经网络与群智能优化方法以高效求解模型。【结果】基于30个客户点、10个配送中心的合成数据与典型静态禁飞区、风场模拟场景的试验表明，在复杂多约束无人机配送调度任务中，与深度强化学习（DRL）方法和基于Transformer规划模型相比，多智能体图神经网络算法展现出更高的解质量，优化后的路径能够充分顺应城市复杂空间结构，合理避让静态禁飞区并兼顾飞行安全，全面降低了配送任务成本。【结论】MASGNN算法在城市复杂场景下展现出高效全局寻优能力，为城市低空物流的调度提供了可靠的智能优化支持。【应用】该优化框架可应用于城市物流、应急物资投送以及多场景协同配送中的智能无人机调度系统设计与实际部署。
Abstract：[Background] Urban multi-depot unmanned aerial vehicle(UAV) delivery systems hold significant promise for intelligent logistics and low-altitude urban economies. However, they encounter complex challenges such as path planning, no-fly zone avoidance, and wind field disturbances,which substantially complicate scheduling optimization. [Objective] This study aims to optimize path planning algorithms for multi-depot UAV delivery under complex constraints, with a focus on minimizing delivery distance and energy consumption while ensuring safe obstacle avoidance.[Methods] A multi-objective mixed-integer programming model is proposed, incorporating delivery demand, flight time windows, no-fly zone constraints, and wind field effect.To solve the model efficiently, we integrate graph neural networks with swarm intelligence optimization techniques. [Results] Experiments using synthetic data from 30 customer locations and 10 depots, combined with simulated no-fly zones and wind fields, demonstrate that the MAS-GNN algorithm outperforms deep reinforcement learning(DRL) and transformer-based methods in both solution quality and stability.The algorithm adapts effectively to complex urban environments, ensure flight safety, avoids restricted areas, and significantly reduces overall delivery costs. [Conclusions] The MAS-GNN algorithm exhibits strong global optimization performance and robustness in complex urban scenarios, offering reliable intelligent scheduling support for urban low-altitude logistics.[Application] This framework is applicable to the design and deployment of intelligent UAV scheduling systems for urban logistics,emergency supply delivery, and multi-scenario collaborative transportation.
[1]陈亚青，张亮，杨志海.无人机路径规划研究现状与发展[J].系统仿真学报，2019, 31(5):827-833.CHEN Yaqing, ZHANG Liang, YANG Zhihai. Research progress and prospect of UAV path planning[J]. Journal of System Simulation, 2019, 31(5):827-833.
[2]XIAO M, HU P, ZOU S, et al. A review of recent advances in smart logistics:From traditional to innovative logistics[J]. Transportation Research Part E:Logistics and Transportation Review, 2023, 172:103099.
[3]缪希仁，李斌，陈军.基于图像识别的无人机电力巡检关键技术[J].电力系统技术，2019, 43(1):286-293.MIAO Xiren, LI Bin, CHEN Jun. Key technologies of UAV power inspection based on image recognition[J].Power System Technology, 2019, 43(1):286-293.
[4]HUANG R, LIU Q, WANG S, et al. A review of dronebased logistics:opportunities, challenges and future trends[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(3):2154-2170.
[5]KIM H C, LEE D Y, CHOI S J, et al. Drone logistics for urban last-mile delivery:a review of challenges and solutions[J]. Journal of Intelligent&Robotic Systems, 2023,109(1):1-19.
[6]CAO T, CHEN Y, GAO Z, et al. Emergency logistics and drone delivery in urban areas:a comprehensive review[J].Computers&Industrial Engineering, 2024, 189:109968.
[7]ZHAO L, REN H, GUO W, et al. Multi-depot UAV routing problem with time windows and heterogeneous fleet[J]. Computers&Industrial Engineering, 2022, 171:108422.
[8]WANG D, LIU K, LIU Y. Multi-UAV coordination for urban logistics delivery:a review and future directions[J].IEEE Transactions on Intelligent Transportation Systems,2023, 24(9):9951-9965.
[9]LI B, WANG X, CAI Z, et al. Dynamic UAV scheduling for last-mile delivery considering time windows and battery constraints[J]. Computers&Industrial Engineering,2020, 148:106716.
[10]MACIAS M, ZHANG G, WANG F, et al. Energy-aware multi-stage UAV routing and battery recharging optimization[J]. IEEE Transactions on Intelligent Transportation Systems, 2023, 24(6):7650-7662.
[11]任璇，王炜，张辉.考虑多作业模式的车辆-无人机协同路径规划问题[J].交通运输系统工程与信息，2021, 21(1):50-58.REN Xuan, WANG Wei, ZHANG Hui. Vehicle-UAV collaborative routing problem considering multiple working modes[J]. Journal of Transportation Systems Engineering and Information Technology, 2021, 21(1):50-58.
[12]SONG Y, JIANG J, WANG Y. Multi-UAV cooperative delivery routing under payload and battery limitations[J]. Journal of Intelligent&Robotic Systems, 2022, 104(3):68-82.
[13]SCHULZ A, GOEBEL C, SCHMIDT M. UAV routing:a survey of optimization problems and algorithms[J]. European Journal of Operational Research, 2020, 280(3):1007-1025.
[14]YIN Y, ZHANG Y, WANG H, et al. A fuzzy chance-constrained approach for electric vehicle routing with hybrid uncertainties[J]. European Journal of Operational Research, 2021, 294(2):643-660.
[15]LI B, WANG C, WU J, et al. Distributed robust optimization model for stochastic inventory routing problem under uncertain demands[J]. Computers&Operations Research, 2020, 117:104886.
[16]GHOSAL A, RAHIMI M, KARIMI I. Chance-constrained capacitated vehicle routing under demand uncertainty[J]. European Journal of Operational Research,2021, 291(2):612-627.
[17]沈燕，卢恺，尹浩东，等.城市飞行汽车停机坪选址问题研究一：基于双层规划的多目标优化方法[J].交通运输工程与信息学报，2025, 23(3):74-87.SHEN Yan, LU Kai, YIN Haodong, et al. Locating urban flying car vertiports part I:a multi-objective optimization method based on bi-level programming[J]. Journal of Transportation Engineering and Information,2025, 23(3):74-87.
[18]曹炳汝，周伟，唐琪.面向多车型动态需求的两阶段车辆路径优化[J].系统仿真学报，2021, 33(8):1591-1599.CAO Bingru, ZHOU Wei, TANG Qi. Two-stage vehicle routing optimization for multi-vehicle dynamic demand[J]. Journal of System Simulation, 2021, 33(8):1591-1599.
[19]范文博，肖杰，吴艺楷，等.低空飞行器运行优化决策方法综述与展望[J].交通运输工程与信息学报，2025,23(3):37-59.FAN Wenbo, XIAO Jie, WU Yikai, et al. A review of optimization methods for low-altitude aircraft operations[J]. Journal of Transportation Engineering and Information, 2025, 23(3):37-59.
[20]丁杰，王迪.基于强化学习的城市低空无人机路径规划[J].交通运输工程与信息学报，2025, 23(2):189-206.DING Jie, WANG Di. Urban low-altitude UAV path planning based on reinforcement learning[J]. Journal of Transportation Engineering and Information, 2025, 23(2):189-206.
[21]WANG C, ZHANG M, LIU S, et al. Graph neural networks for vehicle routing problems:a survey[J]. Swarm and Evolutionary Computation, 2023, 79:101297.
[22]KIM J, SONG Y, LEE J. Multi-agent deep reinforcement learning for UAV path planning in dynamic environments[J]. Expert Systems with Applications, 2023, 227:119934.
[23]ZHANG S, LIU H, LI J, et al. Multi-agent reinforcement learning for UAV swarm control:a survey[J]. Robotics and Autonomous Systems, 2023, 169:103525.
[24]WU S, LI C, HUANG T, et al. Multi-agent path finding based on graph neural networks for large-scale logistics systems[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 253:360401.
基本信息:
DOI：10.19961/j.cnki.1672-4747.2025.06.031
中图分类号:TP183;V355
引用信息:
[1]周星宇,李姝涵,薛锋.基于多智能体图神经网络的城市无人机配送路径优化[J].交通运输工程与信息学报,2025,23(04):50-61.DOI:10.19961/j.cnki.1672-4747.2025.06.031.
基金信息:
国家级大学生创新创业训练计划项目（202210613046）
2025-08-21
2025-08-21
2025-08-21