Combined Location Models

这个提案是我之前由NSERC资助的提案的延续。在过去的5年里，我取得了相当大的进步。***早期的选址模型假设设施可以毫不延迟地为所有客户提供服务，它们永远不会失败，并且可以分别做出库存和营销决策。提出了放宽这些假设的组合区位模型的四个主要领域。***可靠性-位置模型：本研究旨在更好地理解中断如何影响最佳位置决策。我建议6个方向来实现它。由于在大多数应用中，客户需求是连续分布的，而大多数研究假设需求是离散的，因此本研究的重点是研究连续需求的定位问题。这项工作将协助管理人员在预算分配、建造新设施、提高现有设施的可靠性和改进信息系统方面作出更好的决定。***营销-区位模型：我们提出的研究属于设施区位模型和营销扩散模型的交叉。早期的研究通常忽略了产品所获得的市场支持和时间效应，如口碑，而后者往往忽略了产品应该在哪里可以买到，以及旅行距离对需求的影响。主要目标是更好地理解系统的动态，开发新模型并扩展现有模型，以显示将营销扩散模型和位置决策相结合的含义。我建议研究4个方向。***排队-位置模型：在我们关注的这类模型中，客户产生随机的需求流，服务器容量有限，服务时间是随机的，因此到达设施的客户有时必须等待或离开设施。目标是设计一个系统，其中主要决策是：设施的数量、位置、服务能力和客户对设施的分配。我提出了7个旨在帮助服务管理者提高系统质量的研究方向。***库存-定位模型：将传统的供应链优化管理问题和定位问题分开研究。由于这通常会导致低效率和高成本，我们计划将它们联合考虑。主要决策包括找到设施（供应商、零售商和配送中心）的数量、它们的位置和需要订购的库存数量。我提出了5个研究方向，旨在更好地理解联合库存和选址决策。我们还期望分别对这两个领域的方法作出贡献。***我们注意到，本提案中考虑的一些问题，如缩短等待时间（例如，在医疗保健方面）和确定可靠的设施在加拿大很重要。**

This proposal is a continuation of my former proposal funded by NSERC. As will be shown in the proposal I have made considerable progress over the last 5 years.***Early location models assumed that facilities can serve all customers without delay, that they never fail and that decisions on inventory and marketing can be made separately. Four major areas of combined location models that relax these assumptions are proposed.***Reliability-Location Models: This research seeks to gain a better understanding of how disruptions affect optimal location decisions. I suggest 6 directions to achieve it. Since in most applications, customers' demand is distributed continuously while most research assumes discrete demand, a main focus of the proposed research is on location problems with continuous demand. The work will assist managers in making better decisions on budget allocation, on constructing new facilities, on improving reliability of existing facilities and on improving the information system. ***Marketing-Location Models: Our proposed research belongs to the intersection of facility location models and marketing diffusion models. The earlier research typically ignores the marketing support that the product receives and the temporal effects like word-of-mouth while the latter tends to ignore where the product should be available and the effect of travel distances on the demand. The main goal is to understand better the dynamics of the system and to develop new models and extend existing ones in order to show the implications of combining marketing diffusion models and location decisions. I suggest 4 directions of research. ***Queuing-Location Models: In the class of models we focus on, customers generate a stochastic stream of demands, servers have limited capacity, service times are stochastic and therefore customers arriving at the facilities must wait sometimes or leave the facility. The goal is to design a system where the main decisions are: number of facilities, their locations, their service capacities and the assignment of customers to facilities. I suggest 7 directions of research that aim at helping service managers improve system quality. ***Inventory-Location Models: Traditionally optimal management of supply chain problems and location problems were studied separately. As this usually may lead to inefficiencies and high costs, we plan to consider them jointly. The main decisions include finding the number of facilities (suppliers, retailers and distribution centers (DCs)), their locations and the amount of inventory to order. I propose 5 directions of research that aim at gaining a better understanding of joint inventory and location decisions. We also anticipate contribution to the methodologies of both areas separately.***We note that some of the issues considered in this proposal such as reducing waiting times (e.g., in healthcare) and locating reliable facilities are important in Canada. **