Empirical heuristics for improving intermittent demand forecasting

Abstract

Purpose

Intermittent demand appears sporadically, with some time periods not even displaying any demand at all. Even so, such patterns constitute considerable proportions of the total stock in many industrial settings. Forecasting intermittent demand is a rather difficult task but of critical importance for corresponding cost savings. The current study aims to examine the empirical outcomes of three heuristics towards the modification of established intermittent demand forecasting approaches.

Design/methodology/approach

First, optimization of the smoothing parameter used in Croston's approach is empirically explored, in contrast to the use of an a priori fixed value as in earlier studies. Furthermore, the effect of integer rounding of the resulting forecasts is considered. Lastly, the authors evaluate the performance of Theta model as an alternative of SES estimator for extrapolating demand sizes and/or intervals. The proposed heuristics are implemented into the forecasting support system.

Findings

The experiment is performed on 3,000 real intermittent demand series from the automotive industry, while evaluation is made both in terms of bias and accuracy. Results indicate increased forecasting performance.

Originality/value

The current research explores some very simple heuristics which have a positive impact on the accuracy of intermittent demand forecasting approaches. While some of these issues have been partially explored in the past, the current research focuses on a complete in‐depth analysis of easy‐to‐employ modifications to well‐established intermittent demand approaches. By this, the authors enable the application of such heuristics in an industrial environment, which may lead to significant inventory and production cost reductions and other benefits.

Citation

Petropoulos, F., Nikolopoulos, K., Spithourakis, G.P. and Assimakopoulos, V. (2013), "Empirical heuristics for improving intermittent demand forecasting", Industrial Management & Data Systems, Vol. 113 No. 5, pp. 683-696. https://doi.org/10.1108/02635571311324142