Optimization of Raw Material Inventory Control Using the Probabilistic Economic Order Quantity (EOQ) Method

Abstract

Effective raw material inventory management is essential for production continuity in manufacturing companies, yet conventional Economic Order Quantity models assume that demand and lead time are constant and known with certainty. This study applies a probabilistic EOQ approach to optimize raw material inventory management by explicitly incorporating demand and lead-time variability observed over twenty-four months of historical data from a manufacturing company. Descriptive analysis showed that monthly demand averaged 4,123 units with a standard deviation of 506 units, while supplier lead time averaged 8.95 days with a standard deviation of 1.86 days. The deterministic EOQ baseline yielded an order quantity of 4,831 units per replenishment, subsequently extended into a probabilistic framework evaluating five target service levels of 90%, 95%, 97%, 98%, and 99%. Total inventory cost followed a J-shaped pattern across service levels, reaching its minimum of approximately Rp18.09 million per year at the 97% service level, corresponding to a safety stock of 708 units and a reorder point of 1,939 units. Comparison with the company's existing policy showed that although current practice produced a lower nominal cost, it lacked an explicit safety-stock mechanism and had experienced an actual stockout event, indicating that its apparent cost advantage reflects insufficient protection rather than genuine efficiency. Sensitivity analysis confirmed that total cost is highly responsive to ordering and holding cost variation, whereas the optimal order quantity remains stable. These findings demonstrate that probabilistic EOQ modeling offers a more transparent, risk-aware basis for raw material replenishment decisions than deterministic or intuition-based approaches.