Analysis of uncertainty levels using dynamic programming for an integrated production and inventory management problem of a perishable product

In this paper we present a sensitivity analysis in relation to demand uncertainty considering a dynamic programming approach for an integrated production and inventory management problem. The model considers restrictions related to production, storage and flow of materials between production and availability of products to retailer’s customers. We propose a rolling horizon strategy for generating plans for a perishable product. At each decision-making stage, there is adaptation of expected demand levels and re-planning in the face of the fulfillment of initially uncertain demands. We carried out an investigation of the relationship between different levels of uncertainty and the indicators: destroyed items, service level and expected profits and compared the results with the results obtained considering a static approach. The results highlight a large performance gain in the dynamic approach, compared to the static approach, and make it possible to estimate the gains obtained when investing in more accurate demand forecasts.