Collaborative Research: Integrated Supply Chain Design for Customer Service, Robustness and Information Systems

Northwestern University, Evanston IL

Investigators

Abstract

Inventory management has become a major concern in the service industries. Heightened customer service expectations and reduced product life spans have complicated inventory decision-making. Firms must decide where to stock inventory and how much inventory to maintain to satisfy customer demands. We will develop a new integrated facility location and inventory models that will allow managers to better decide (1) how many facilities to have, (2) where those stocking locations should be, (3) how much inventory to stock at each site, (4) how often to reorder inventory, and (5) which customers to assign to each site to minimize overall costs while satisfying specified customer service requirements. The research will build on earlier work of the PIs in which we have developed an elementary location/inventory model. We will enhance the modeling approach in a number of different ways to allow the model to capture actual location/inventory problems more realistically. In particular, we will incorporate (1) capacity restrictions at the distribution centers (DCs) and in the various transport channels, (2) more realistic local delivery routing from the distribution center(s) to the retailers including less-than-truckload (LTL) shipments, (3) multiple shipment modes from suppliers to DCs, and (4) multiple classes of inventory and multiple customer types. We will develop models that consider customer service requirements and that optimize service subject to resource constraints. We will then develop multi-objective models that explore tradeoffs between service and cost. Another significant task will be incorporating future scenarios in location/inventory models to identify solutions that are robust with respect to uncertainty about the future. We also will develop a new class of location models based on concepts of reliability. A solution is deemed reliable if the system can continue to function adequately even if parts of the system (e.g., DCs or network links) are removed. Finally, we will explore the incorporation of information sources into integrated facility location/inventory models. A variety of solution procedures will be employed for the models to be developed including optimal, non-linear integer programming based algorithms when possible and heuristic procedures when optimal algorithms cannot be devised.