Improving The Control of Inventory Management Systems Using Robust Estimators in the Presence of Outliers

Abstract

Inventory management systems aim to control inventory levels in the best way while reducing costs to a minimum. However, inventory management faces many challenges that lead to the deficit or increase in storage costs, and among these challenges is the presence of outliers in demand data, therefore this paper seeks to building a new model that uses robust estimators (Median and Median Absolute Deviation) instead of classical estimators that are highly sensitive to outliers, and examining this model on real data for the demand for raw materials used in the cement industry in one of Iraq's factories. The proposed model was able to reduce the deficit ratio by 96% and contributed to reducing the total cost of storage by 75%, in addition to that, the proposed model contributed to reducing the need for safety stock by 30%.