人机协同：无人机与骑手混合配送模式的效率、成本与用户体验研究

Human-Machine Collaboration: A Study on the Efficiency, Cost, and User Experience of a Hybrid Drone-Rider Delivery Model

摘要: 即时配送“最后一公里”面临效率瓶颈、成本攀升与体验难平衡的挑战。无人机与骑手协同配送被视为破局关键，但其系统性价值有待量化验证。本研究构建了一个整合运营效率、经济成本与用户体验的多目标动态调度模型，并设计了自适应大邻域搜索(ALNS)算法进行求解。通过仿真实验，对比分析了纯骑手模式、静态协同规则与人机协同优化模式的性能。数值实验表明，人机协同优化模式能实现显著的协同效应：与纯骑手模式相比，总成本降低约12.9%，订单准时率提升至94.7%，平均配送时长缩短17.4%。协同效益受无人机单位成本、订单密度及区域特征显著影响，存在关键成本阈值。本研究首次将“用户体验”以非线性延误惩罚函数形式内生于协同调度模型，为平台提供了可量化权衡“效率–成本–体验”的战略决策工具，并提出了优先在高潜力区域部署的精准投资建议。

Abstract: The “last-mile” instant delivery faces challenges of efficiency bottlenecks, rising costs, and difficulty in balancing user experience. The collaboration between drones and riders is seen as a key solution, yet its systematic value needs quantitative verification. This study constructs a multi-objective dynamic scheduling model integrating operational efficiency, economic cost, and user experience, and designs an Adaptive Large Neighborhood Search (ALNS) algorithm for the solution. Through simulation experiments, the performance of pure rider mode, static collaboration rules, and human-machine collaborative optimization mode is compared and analyzed. Numerical experiments show that the optimized human-machine collaboration model can achieve significant synergistic effects: compared with the pure rider mode, the total cost is reduced by about 12.9%, the on-time delivery rate is increased to 94.7%, and the average delivery time is shortened by 17.4%. The benefits of collaboration are significantly affected by drone unit cost, order density, and regional characteristics, with a critical cost threshold existing. This study is the first to incorporate “user experience” into the collaborative scheduling model in the form of a non-linear delay penalty function, providing platforms with a strategic decision-making tool for quantitatively trading off “efficiency-cost-experience”. It also proposes precise investment suggestions for priority deployment in high-potential areas.