题名: | 多卡车-多异构无人机协同配送路径优化研究 |
作者: | |
学号: | SX2207063 |
保密级别: | 公开 |
语种: | chi |
学科代码: | 082303 |
学科: | 工学 - 交通运输工程 - 交通运输规划与管理 |
学生类型: | 硕士 |
学位: | 工学硕士 |
学校: | 南京航空航天大学 |
院系: | |
专业: | |
导师姓名: | |
导师单位: | |
完成日期: | 2024-12-18 |
答辩日期: | 2025-03-14 |
外文题名: |
Collaborative distribution route planning research for multi-truck and multi-heterogeneous drones |
关键词: | 多卡车-异构无人机协同配送 ; 混合整数规划 ; 分布式鲁棒优化 ; 变邻域搜索 ; 模拟退火算法 |
外文关键词: | Multi-truck and heterogeneous drone collaborative delivery ; Mixed-integer programming ; Distributionally robust optimization ; Variable neighborhood search ; Simulated annealing algorithm |
摘要: |
电商行业的迅速发展使传统“最后一公里”配送面临挑战,无人机技术进步为卡车-无人机协同配送系统带来新的可能性。因此,研究多卡车-多异构无人机协同配送路径优化问题具有重要意义。 首先,本文在VRP-D模型基础上,考虑异构无人机、多访问、无人机能量消耗与载重、飞行时间的非线性关系,构建了以最小化完工时间为目标的混合整数规划模型。同时,为应对配送环境的不确定性,以行驶时间为不确定参数,通过分布式鲁棒优化方法对模型进行扩展。其次,针对卡车和无人机的两级路径特性,设计了两阶段初始解构造方法,并提出混合变邻域搜索与模拟退火算法(VNS-SA)进行求解优化。算法设计了四种邻域变换算子以扩大搜索范围,使用模拟退火算法避免陷入局部最优。 实验验证了模型和算法的有效性。在小规模算例中,与Gurobi求解器相比,随着规模增大VNS-SA求解速度和质量较好。在大规模算例中,较传统卡车配送模式显著缩短了配送时间。在配送区域较小时,无人机能量限制较低,完工时间更短;异构无人机能够各自发挥特点,进一步提升配送效果。此外,灵敏度分析显示,无人机性能对配送效率影响显著,但单一性能提升的作用有限。算子实验结果表明,无人机路径生成算子对算法性能的贡献最大,四个算子的协同作用提升了解的质量。最后,基于亚马逊“最后一公里”配送数据集,讨论了不同电池更换假设及不确定行程时间对配送效果的影响。本研究为多卡车-异构无人机协同配送路径优化提供了理论支持与技术实现方案。 |
外摘要要: |
In recent years, the rapid growth of the e-commerce industry has introduced substantial challenges to traditional last-mile delivery systems. The advent of drone-assisted delivery technology presents promising opportunities for truck-drone collaborative logistics. Consequently, optimizing multi-truck and multi-heterogeneous drone collaborative delivery systems holds significant importance for addressing these challenges. This study develops a mixed-integer programming model based on the VRP-D framework, incorporating key factors such as heterogeneous drones, multiple visits, and the nonlinear relationships among drone energy consumption, payload, and flight time. To address uncertainties in the delivery environment, travel times for both trucks and drones are treated as uncertain parameters. A distributionally robust optimization approach is applied to extend the deterministic model, enhancing its adaptability to real-world conditions. Given the two-tiered routing characteristics of trucks and drones, the study proposes a two-stage initial solution construction method and a Hybrid Variable Neighborhood Search and Simulated Annealing (VNS-SA) algorithm for solution optimization. The algorithm integrates four types of neighborhood transformation operators to expand the search space, while simulated annealing is employed to mitigate the risk of being trapped in local optima. The experiments validate the effectiveness of the proposed model and algorithm. In small-scale instances, the VNS-SA algorithm outperforms the Gurobi solver in both solution speed and solution quality as the case scale increases. In large-scale instances, the truck-drone collaborative delivery mode significantly reduces delivery time compared to the traditional truck-only mode. The experiments also demonstrate that in smaller delivery regions, the energy constraints of drones are less restrictive, resulting in shorter completion times. Heterogeneous drones can leverage their respective characteristics to further enhance delivery performance. Moreover, sensitivity analysis reveals that drone performance significantly impacts delivery efficiency, although the benefits of improving a single performance metric are limited. Operators experiment results indicate that the drone route generation operator contributes the most to algorithm performance, while the collaborative effect of all four operators improves solution quality. Finally, based on the Amazon last-mile delivery dataset, the impact of different battery replacement assumptions and uncertain travel times on delivery performance is discussed. This study provides theoretical support and practical implementation strategies for optimizing the routing of multi-truck and heterogeneous drone collaborative delivery systems. |
参考文献: |
[28]罗志浩. 车载无人机双层协同路径规划问题研究[D]. 国防科技大学, 2017. [37]马华伟,马凯,郭君.考虑多投递的带无人机车辆路径规划问题研究[J].计算机工程,2022,48(08): [58]Ponza A. Optimization of drone-assisted parcel delivery[J]. 2016. |
中图分类号: | U491 |
馆藏号: | 2025-007-0037 |
开放日期: | 2025-09-27 |