Logistics and Transportation in Supply Chains

装运合并(SCL)是将几个小货物有目的地组合在一起，作为一批较大的货物一起发送。虽然订单交付的延迟可能会产生成本，但SCL可以在运输方面产生显著的规模经济，从而降低整体系统成本。如果注意到最长保存时间(合并周期的长度)，则向货物接收方提供的服务不会降级。托运人希望选择适当的集结政策，以确定终止堆积过程和分派货物的适当时机。我们的新工作旨在提出和研究比单纯基于货物总重量和/或最长等待时间的典型策略更一般的策略。我们的新模型能够记录更多关于订单积累过程的信息(例如，单个订单的延迟)，从而允许考虑依赖于这些信息的政策。我们还将通过考虑超越通常假设的泊松分布的马尔可夫到达过程，以及通过纳入连续订单到达与其权重之间的潜在相关性来提高模型的准确性。继续我们最近在矩阵分析方法上的成功，我们将开发有效的求解技术来获得准确的结果。我们将采用更现实的成本结构：延迟惩罚成本(客户对等待的非效用)在等待时间内严格增加。最终目标是确定一般情况下最优策略的形式，并构造有效的算法来确定最优策略的参数，对于涉及运输成本和/或延迟违约率可能不同的多个产品的订单的合并过程。**批量问题寻求确定单个或多个项目的时间和补给量，以在满足需求的同时最小化总成本。基于客户订单的需求可以通过生产和分配来满足，当这些项目被分组到系列中时。通过家庭的联合协调认识到在运输(如在SCL)或在供应链的制造部分共享资源。我们对几个多产品族的批量问题的研究在生产环境中是新的。这项工作还将有助于对从共同供应商订购的产品进行协调库存控制。我们将为多家族批量问题探索各种强大的重构方法，并测试新的启发式算法以加快求解时间，保持解的质量。**库存路线问题既涉及车辆路线(客户通过卡车送货的排序)，也涉及供应商为其产品正在送货的客户管理库存。我们的研究允许，在有利的情况下，开发间接路线。这些服务是为通过给定的交叉码头(CD)接收货物的逻辑分组的客户集服务的。在车辆容量允许的情况下，路线上可以有尽可能多的客户。同时考虑了客户的库存持有成本和管道库存。这一主题以一种不平凡的方式扩展了申请人以前关于VMI(供应商管理库存)和交叉对接的出版物。我们希望为一个包含多个CD的系统获得普遍的、可证明的接近最优的结果和有效的求解算法，在该系统中，来自多个供应商的不同产品由私人车队通过路线交付给不同的客户。这项工作将代表最早的此类模型之一。

SHIPMENT CONSOLIDATION (SCL) is the purposeful combination of several small loads, for dispatch together, as a single larger shipment. Although delay of order deliveries may incur a cost, SCL can lead to significant economies of scale in transportation, thus reducing the overall system cost. Service to the receiver of the goods need not be degraded, if attention is paid to the maximum holding time (length of the consolidation cycle). The shipper wishes to select an appropriate consolidation policy that determines the right moment to terminate the accumulation process and dispatch a load. Our new work aims to propose and study policies that are more general than the typical ones that are based solely upon the aggregate shipment weight and/or the maximum holding time. Our new models are capable of recording more information about the order-accumulation process (e.g. delays of individual orders), thus permitting consideration of policies dependent on such information. We will also improve the accuracy of our models by considering Markovian arrival processes that go beyond the typically-assumed Poisson distribution, and by incorporating potential correlations between the arrivals of consecutive orders and their weights. Continuing our recent success with the Matrix-Analytic method, we will develop efficient solution techniques to obtain accurate results. We will employ a more realistic cost structure: a delay penalty cost (customer disutility of waiting) that is strictly increasing in that waiting time. The ultimate goals are to identify the form of the optimal policy for the general case, and to construct efficient algorithms to determine the parameters of that optimal policy, for a consolidation process involving orders of multiple products whose transport costs and/or delay-penalty rates may differ.**LOT-SIZING PROBLEMS seek to determine the timing and replenishment quantities, of a single item or of several items, to minimize total costs while meeting demand requirements. Demand, based on orders from customers, can be satisfied through the production and distribution when those items are grouped into families. The joint coordination via families recognizes the sharing of a resource in transportation (as in SCL) or in the manufacturing portion of the supply chain. Our research in the study of lot-sizing of several multi-product families is new in the production context. This work will also shed light on coordinated inventory control of products ordered from a common supplier. We will explore various strong reformulations for the multi-family lot-sizing problem, and test new heuristics to speed up solution times, maintaining solution quality. * *THE INVENTORY-ROUTING PROBLEM concerns both vehicle routing (the sequencing of customers in making deliveries by truck) and the management by the vendor of stocks for those customers whose products are being delivered. Our research allows, when advantageous, the development of indirect routes. These serve logically-grouped sets of customers that receive goods through a given cross-dock (CD). There can be as many customers on the routes as permitted by vehicle capacity. Both the inventory holding costs at customers and the pipeline stocks are considered. This theme extends, in a non-trivial way, the applicant's previous publications on VMI (Vendor Managed Inventory) and cross-docking. We hope to obtain general, provably near-optimal results, and an efficient solution algorithm, for a system containing several CDs in which distinct products from multiple suppliers are delivered by a private fleet on routes to various customers. This work will represent one of the very first such models.