A simulation-based approach to find the optimal order size for seasonal products under uncertainty on forecast parameters

Muñoz D.F. · Expert Systems with Applications (2026) · DOI: 10.1016/j.eswa.2026.132807

Resumen

We discuss how a general Bayesian approach can be applied to determine the optimal order size for seasonal or new products, so that it can be applied using different optimality criteria to incorporate the uncertainty in the parameters of the demand forecast. The application of this methodology is illustrated through two relevant criteria for determining the optimal order size: the well-known newsboy problem and the capacity index criterion proposed by Su and Pearn (2011). Experiments are conducted with a forecast model that has an analytical solution for the posterior distribution, and experimental results to illustrate the estimation of the optimal order size using Monte Carlo simulation are discussed. Monte Carlo simulation is appealing when the complexity of the forecast model does not allow us to obtain the optimal order sizes through an analytical solution or to use a more efficient numerical algorithm.

Instituciones: Universidad Nacional Agraria La Molina