Quantile Forecasting in Operational Planning and Inventory Management – an Initial Empirical Verification

Abstract

In the paper we present our initial results of an empirical verification of different methodologies of quantile forecasting used in operational management to calculate the re-order point or order-up-to level as well as the optimal order quantity according to the newsvendor model. The comparison encompasses 26 procedures including quantile regression, the basic bootstrap method and popular textbook formulas. Our results, obtained on the base of 30 time series concerning such diversified phenomena as supermarket sales, passenger transport and water and gas demand, point to the usefulness of regression medians, regression quantiles, bootstrap methods and the procedures available in the SAP ERP system.