To aggregate or not to aggregate: Forecasting of finite autocorrelated demand

Temporal aggregation is an intuitively appealing approach to deal with demand uncertainty. There are two types of temporal aggregation: non-overlapping and overlapping. Most the supply chain forecasting literature has focused so far on former there no research that analyses latter for auto-correlated demands. In addition, most analytical to-date assumes infinite series’ lengths whereas, in practice, based finite histories. The length history important determinant comparative performance approaches but not been given sufficient attention literature. this article, we examine effectiveness demand. We do by means study forecast accuracy non-aggregation mean-squared error. complement a numerical analysis explore impact parameters series (comparative) performance. also conduct empirical evaluation validate results using monthly time M4-competition dataset. find degree auto-correlation, horizon be determinants accuracy. discuss merits each highlight their implications real-world practices.