Controle de reposição e estoques de um Item com demanda incerta via programação estocástica (Replacement control and inventory of an item with uncertain demand via stochastic programming)

In this paper, an inventory and replenishment control system is determined for a logistics network with one layer, assuming a single item with uncertain demand, via stochastic programming. The periodicity of replenishment orders as well as the variable replenishment quantities are determined using mixed integer linear programming, which takes into account the uncertain nature of the levels of demand for the item through a finite set of scenarios. In this sense, a sampling system based on Monte Carlo simulation, known in literature as "sample average approximation" (SAA), was used. Numerical results are presented with the aid package AIMMS, and suggest that the obtained minimum costs for known instances of the literature, via stochastic programming, show an error of 1% approximately, when compared to the optimal costs, even with a small number scenarios.