Approximations for the relative outdating of perishable products by combining stochastic modeling, simulation and regression modeling

In this paper we derive approximations for the expected amount of outdating (waste) in an inventory system with perishable products, periodic review, positive lead time, fixed case pack sizes and a First-In-First-Out withdrawal policy. Demand is discrete and stochastic. Demand which is not met from inventory on hand will be lost. The approximations can be used to evaluate how the amount of outdating can be reduced by changing the safety stock and/or the product and demand characteristics. We first derive approximations in an analytical way and then use regression to improve these approximations. The final approximation is very fast and performs very well: the mean and standard deviation of the approximation error are equal to −0.01% and 0.78%. Moreover, the approximation is applicable in situations with parameter settings which are frequently encountered in practice, and for which no other approximations have been derived in the literature before.