Periodic Review for Perishable Products Under Service Level Constraints : Optimization, Sensitivity Analysis, and Predictive Modelling

A multi-period inventory management model for perishables is developed using a simulation approach. Single-product scenario and periodic review policy are assumed, considering uniform demand, non-null deterministic lead time, fixed shelf life and service level constraints.Simulation is used to determine the optimal operating leverages, namely order-up-to-level (OUTL) and reorder period (DT), which minimize the total cost under service level restrictions. Seven input factors related to managerial aspects and product characteristics, and three output responses are considered during the simulation campaign based on Design of Experiments. Analysis of Variance is performed to identify the significant parameters. The results show that the optimal operating leverages depend on all the input factors, except for the stock-out cost. Furthermore, the shelf life results to be significant only below a given (critical) value, that depends on the specific context; for higher shelf life values the products could be treated as non-perishable, from an inventory management point of view.The significant parameters are then embodied in predictive polynomial models, according to Response Surface Methodology, for the estimation of the optimal levels of OUTL and DT, able to meet the service level constraints at the minimum cost. These models present various advantages for industrial decision-makers, thanks to the capability of providing instantaneous estimates of optimal OUTL and DT, and to their suitability for being implemented in advanced management tools, including the development of digital twins. Finally, the models are validated by simulating additional case studies, whose outcomes are compared with those returned by the estimate