Inventory optimization under demand uncertainty: algorithmic approaches

The purpose of this master thesis is to suggest algorithmic methods that can be used to solve real-life inventory management problems. The approaches should be scalable in order to be adapted to specific contexts and fit to many firms’ needs. Moreover, the methods must consider demand uncertainty to propose robust solutions. We focus our research on inventory management problems under periodic review systems. For these problems, there are several replenishment strategies and different algorithmic approaches. We replicate two methods from the literature and we develop two other methods using a dynamic programming formulation. We confront the four methods to a test bed and we carry out a sensitivity analysis. This experiment is made to assess the performance of the methods under different settings. To assess the scalability of the methods, we attempt to extend them, and we compare the performances of the extended methods on common test beds.Master [120] en Ingénieur de gestion, Université catholique de Louvain, 201