New Algorithms for Distribution Systems under Demand and Supply Uncertainties
Abstract
|
Distribution systems are among the most difficult network topologies for inventory optimization. The optimal inventory and allocation policies for these systems are unknown. In this talk, we present simple algorithms for optimizing base-stock levels in multi-echelon distribution systems subject to demand or supply uncertainties. Initially, we consider a system with random customer demands and propose a heuristic procedure to approximate the order-up-to levels of all the locations in the system. Our Decomposition-Aggregation (DA) heuristic is both more accurate and less computationally intensive than approaches that have been recently proposed in the literature. Because of its simplicity and effectiveness, the DA heuristic provides an opportunity to address managerial questions like the responses of the locations in the system to changes in the supply chain configuration. In the second part of the talk, we present two-echelon distribution systems subject to supply disruptions. We consider a range of assumptions and propose algorithms to find the optimal or near-optimal stocking levels of all the locations in the system. We show how supply disruptions in different parts of the network affect inventory decisions and propose strategies that firms can use to mitigate the impacts of these disruptions. We conclude that ignoring supply disruptions closer to customers has more negative effects than ignoring the disruptions in the other parts of the chain. In addition, we show that firms should focus more on reducing the duration of disruptions than on reducing their probability of occurrence. |
| Contact information: |
| Dr. K.J. Roodbergen |
