报告时间：2026年9月30日（星期三）9:30-12:00
报告地点：格物楼517
报 告 人：张真真 教授
工作单位：同济大学
举办单位：汽车与交通工程学院
报告简介：
In many pickup logistics systems, vehicles must collect heterogeneous items whose exact composition is unknown until arrival, while loading decisions are sequential and irreversible due to reloading bans. In such settings, feasibility cannot be characterized by scalar capacity measures, but depends on the geometric and sequential structure of loading decisions. Thus, we study a stochastic vehicle routing problem in which loading feasibility is path-dependent and nonmonotone, fundamentally differing from classical stochastic VRPs. We formulate the Three-Dimensional Loading Pickup Capacitated Vehicle Routing Problem with Stochastic Demands (3L-PCVRP-SD) under an a priori framework with detour-to-depot recourse. Demand uncertainty is represented through data-driven scenarios with clustering-based reduction, and routing decisions are evaluated via recourse simulation under loading constraints. Computational results show that explicitly modeling loading feasibility leads to substantial reductions in recourse costs. They further reveal how path-dependent loading constraints reshape the trade-offs between routing cost, loading flexibility, and uncertainty.
报告人简介：
张真真，同济大学经济与管理学院长聘教授、博士生导师。长期致力于大规模整数规划和不确定优化的理论研究与算法设计，及在物流与运输规划、智能制造等方面的应用。目前已发表高质量论文30余篇，包括Operations Research、INFORMS Journal on Computing、Transportation Science、Transportation Research Part B、NeurIPS等，主持国家自然科学基金青年项目及优秀青年项目、上海市人才项目和华为、中远海运科研课题各1项，作为项目骨干参与创新研究群体和重点项目各1项。
