This project aims to develop methods for effective inventory management for supply chains with general topologies without making demand distribution assumptions. Although there is a rich literature on inventory management, this literature generally assumes that the distribution of demand as well as the distribution of production capacity per period is known by the manager. However, in practice, many managers do not have detailed distributional information on demand and capacity. We aim to develop methods for effective inventory control when such distributional knowledge does not exist. We aim to start our research by exploring relatively simpler topologies such as serial systems, two or three retailers with transshipment among them etc. Once we have developed and tested our methods on these simpler topologies, we will extend our analysis to more complex and general topologies where we will develop heuristics and bounds. Our research should contribute to the supply chain inventory management literature as well as have practical impact as it can be used by managers making such decisions on a daily basis.
$290,000 grant from the National Science Foundation
Presented at the INFORMS Annual Conference 2015, Philadelphia, PA
Presented at the INFORMS International 2016, Hawaii
Presented at the Georgia Institute of Technology, ISyE Colloquium Seminar Series, Spring 2016, Atlanta, GA.
Working Paper, University of Michigan, 2016
Published in Operations Research, 2016