基于单边组合VCG拍卖的城市物流联动机制及算法研究

This project is among the first that proposes an efficient auction mechanism for the city logistics synchronization (CLS) problem, which aims to capture both logistics punctuality and simultaneity in a city or region. The main motivation of CLS is that if the delay has already occurred or will occur, customers tend to pursue the simultaneity. We will develop the one-sided Vickrey-Clarke-Groves (O-VCG) auction for the CLS problem. The 3PL company is the auctioneer, while the logistics service purchasers are bidders. The proposed O-VCG auction will realize incentive compatibility (on the buy side), approximately allocative efficiency, budget balance, and individual rationality. The vehicle routing problem (VRP) faced by the 3PL company will be formulated as the lane covering problem with CLS requirements. Three effective heuristics are developed: Merge, Exchange, and Mutate. Our basic computational results show that the three operators are effective but sensitive to the bid duration. A hybrid operator significantly outperforms each individual operator. We also numerically analyze the impacts of five key factors: the strategic behavior of the 3PL company, flexible due dates, the maximum bid duration, the radius of a city or region, and the number of depots. Based on numerical analysis and case study, this project will obtain several important managerial implications for the development of smart city logistics platform and the operations management problems faced by logistics enterprises. This project will not only extend the traditional and conceptual CLS mechanism to the theoretical and computational stages, but provide multiple effective heuristics for the VRPs with CLS requirements.

为解决城市物流联动（CLS）问题，提出一种有效的拍卖机制，实现一个城市或地区物流的准时性和同时性。在延迟已发生或将要发生的情况下，顾客更倾向于追求物流的同时性。3PL公司为拍卖商，物流服务需求方为投标者。提出单边的VCG拍卖机制，旨在实现激励兼容、配置效率、平衡预算、及个体理性。针对带CLS约束的车道覆盖问题，提出三种启发式算法：合并，交换及突变。基本测试显示，三种运算都是有效的，但对投标时间均比较敏感；混合运算明显优于单个运算。用数值分析研究5个关键因素对个体和整体利益的影响：3PL公司的投机行为、可变的交货到期日、最大投标时间槽、一个城市的半径，以及车库的数量。结合数值和案例研究，针对智能城市物流平台建设和物流企业运营管理等问题，充分提炼若干个管理学意义。本研究不仅将长期以来停留在概念层面的CLS机制设计方法推向深入研究阶段，也为带CLS约束的车辆路径问题提供了多种有效的优化算法。

为解决城市或跨区域物流联动问题，提出一种有效的拍卖机制，实现一个城市或区域物流的准时性和同时性。在延迟已发生或将要发生的情况下，顾客更倾向于追求物流的同时性。3PL公司为拍卖商，物流服务需求方为投标者。提出单边的VCG拍卖机制，实现了激励兼容、配置效率、平衡预算、及个体理性。首先，针对一般性的城市物流联动（CLS）机制中的组合优化问题，提出四种启发式算法：合并，交换、突变、及混合。其次，针对城市垃圾清运联动（WCS）问题，提出面向零担物流的VCG拍卖机制，并设计混合变邻域-禁忌搜索（VNTS）算法以求解双层组合优化问题，其中关于上层资源配置问题，构建有效的变邻域搜索（VNS）算法用于寻找可行的资源配置解，而关于下层车辆路径规划问题，构建高效的禁忌搜索（TS）算法用于寻找最小成本的清运路径解。紧接着，针对跨区域物流联动（RLS）机制中的组合优化问题，构建了一系列高效的群体智能优化算法，包括人工蜂群（ABC）算法、具有全局拓扑结构的粒子群（Global PSO）算法、具有局部拓扑结构的粒子群（Local PSO）算法、及具有新型信息学习机制的高级跟踪人工蜂群（STABC）算法。结合数值实验和案例研究，针对智能城市物流平台建设和物流企业运营管理等问题，进行大量的敏感性分析，增强了城市物流联动拍卖机制的实际可行性。本研究不仅将长期以来停留在概念层面的物流联动机制设计方法推向深入研究阶段，也为带物流联动约束的车辆路径问题提供了多种有效的优化算法。最后，本项目对城市停车位的共享、分配与定价也进行了深入研究，取得了较为丰硕的成果，为城市停车位管理机制的建立提供了理论与应用支撑。