Inventory Control of Steel Materials in Shipyard Company Using the Continuous Review Method

Abstract

Ship repair activities are characterized by highly uncertain material requirements because the scope of work is often finalized only after on-dock inspection. This uncertainty frequently leads to material shortages, project delays, and increased operational costs, particularly in replating processes where steel is the primary material. This study aims to develop an improved inventory control policy by integrating demand forecasting and continuous review inventory modeling. An ABC classification was first conducted to identify high-value materials requiring strict control, with Category A items selected for further analysis. Demand forecasting was performed using Exponential Smoothing and Long Short-Term Memory (LSTM) models, and forecasting accuracy was evaluated using Mean Absolute Deviation (MAD) and Mean Squared Error (MSE). The method with the lowest error was then incorporated into a Continuous Review (Q) Model with Backorders to determine optimal order quantities, safety stock, and reorder points. A simulation comparing the proposed policy with the company’s existing practice shows that the proposed approach reduces total inventory-related costs by IDR 2,671,910,250.20, equivalent to a 7.645% cost saving, while improving material availability and reducing shortage risks. The findings demonstrate that combining machine-learning-based forecasting with probabilistic inventory control can significantly enhance inventory performance in project-based industries with fluctuating demand, such as ship repair.