In this work the problem of selective backhauls in the transport of fresh products is investigated for the case of a large Belgian grocery retailer. Explicit measures for estimating the risk involved in certain route - vendor backhauling combinations, which emerge from the uncertainties involved in over road transport and unforeseen waiting and loading/unloading times at the depot or stores, are constructed. Two different models are proposed: an integer linear program with chance constraints, and a stochastic linear program. The chance constraints in the first model are based on the intuition and experience of the people in planning and dispatching. In the second model, the balance between risk and potential profit is directly incorporated into the objective function. For this study, the largest of all transport companies working for the studied retailer is considered for reference and data. The stochastic linear program proved to be superior. Moreover, we demonstrate that, even if only considering a small part of the fleet, the potential profits and reduction in empty kilometers due to the selected backhauls are considerable.

在这项工作中，以比利时一家大型杂货零售商为例，研究了新鲜产品运输中的选择性回程问题。建立了明确的措施来估计某些路线中涉及的风险 - 供应商回程组合，这些风险是由于公路运输的不确定性以及在仓库或商店不可预见的等待和装卸时间而产生的。提出了两种不同的模型：具有机会约束的整数线性规划和随机线性规划。第一个模型中的机会约束基于规划和调度人员的直觉和经验。在第二个模型中，风险和潜在利润之间的平衡直接纳入目标函数。在本研究中，考虑为研究零售商工作的所有运输公司中最大的一家作为参考和数据。随机线性规划被证明是优越的。此外，我们证明，即使只考虑一小部分机队，由于选定的回程而带来的潜在利润和空公里数的减少也是相当可观的。