Adaptive optimization approach for production and distribution planning of perishable food products under demand uncertainty

Abstract Management of production and distribution planning of perishable food products is crucial due to their low-profit margin and the increased environmental costs of manufacturing. The production planning of perishable products is challenging as it combines multiple factors such as temperature tracking of products, sequence-dependent facility setup cost, and uncertain demand in one setting. This paper studies the production and distribution planning problem for perishable food products and unifies the mentioned factors in an optimization framework. We propose an adaptive optimization approach to address the uncertainty in demand, providing flexible optimization approaches for a multi-period planning horizon. The first approach is adjustable robust optimization that generates a resilient and Pareto-efficient production and distribution plan to tackle demand uncertainty avoiding over-conservative solutions via decision rules. The second is the folding horizon, which re-optimizes production and distribution plans based on the realized demands over time. We assess the efficiency of the adaptive approach through extensive Monte Carlo simulation experiments. Furthermore, we perform a real-case adoption study on the production planning of a dairy factory to assess the applicability of our model and solution approach in real-world instances. According to the results, the adjustable approach and the folding horizon approach improve the objective function value by 4–14% and 5–28% for that of the orthodox robust model. The numerical results also show that the adjustable robust approach is always resistant to uncertainty, while the percentage of unmet demand for the deterministic model can reach as high as 18%.