面向电商即时履约的无人配送车动态路径优化与多车协同调度研究

Dynamic Path Optimization and Multi-Vehicle Cooperative Scheduling for Unmanned Delivery Vehicles in E-Commerce Real-Time Fullfillment Scenarios

摘要: 本文针对电商即时履约场景下无人配送车的动态路径优化与多车协同调度问题展开研究。结合上海典型商圈真实路网与模拟订单数据，建立了基于单轮“一车一单”约束的简化PDPTW模型，并将其严格简化为可由匈牙利算法求解的最小权匹配问题。同时，设计工程化蚁群算法与融合策略，实现全局最优与局部启发的平衡。结果表明，匈牙利与融合方案较静态基线在配送距离与能耗上平均降低约20%，在履约总时长上提升约5%，并保持高分配率与稳定性。本文的研究在特定约束条件下，将复杂的动态配送问题转化为高效求解的匹配模型，为生鲜电商和商超即时配送的无人车调度提供可行方法与数据支撑。

Abstract: This paper focuses on the dynamic path optimization and multi-vehicle collaborative scheduling of unmanned delivery vehicles in the context of e-commerce’s immediate fulfillment. By integrating the real road network of typical business districts in Shanghai and simulated order data, a simplified PDPTW model based on the single-round “one vehicle, one order” constraint was established and strictly simplified into a minimum weight matching problem solvable by the Hungarian algorithm. Meanwhile, an engineering ant colony algorithm and a fusion strategy were designed to achieve a balance between global optimality and local heuristics. The results show that the Hungarian and fusion schemes reduced the average delivery distance and energy consumption by approximately 20% compared to the static baseline, improved the total fulfillment time by about 5%, and maintained a high allocation rate and stability. This research, under specific constraints, transforms the complex dynamic delivery problem into an efficiently solvable matching model, providing feasible methods and data support for the unmanned vehicle scheduling in fresh food e-commerce and supermarket immediate delivery.