| Summary: | The aim of this study is to develop the inventory planning system of a Portuguese tyre retailer based on forecasting sales models. Using sales history up to 2020, tyres were grouped into three levels of sales aggregation and different quantitative forecasting models were applied. The comparison of these models resorted to various evaluation measures to choose the most suitable one for each group. The study shows that for items with sales grouped monthly and for items with sales grouped by semester, Holt’s method had a better performance on determining sales forecasts, while for tyres with sales grouped quarterly, it was Croston’s method that stood out. The inventory policy outlined for each group of items reflects the results of the forecasted demand, and the review period depends on the sales group under analysis. In agreement with previous studies, the usefulness of statistical methods is corroborated. Additionally, the advantage of combining the said methods proved helpful, particularly as a starting point for tyre retail inventory planning.
|
|---|