AN OPTIMIZE SHIPMENT DEMURRAGE USING MULTIVARIATE ANALYSIS AND ARTIFICIAL NEURAL NETWORK FORECASTING MODEL

In 2019, XYZ companies suffered a payment penalty, demurrage, as a consequence of the delay in shipping their bulk cargo caused by product shortages. Demurrage donates high financial loss to the company; therefore, an optimization model is proposed to prevent it. This study aims to examine the causes of product shortages during shipment, which affect in demurrage then optimize to avoid the financial losses. This research conducted a multivariate statistical analysis to observe factors affected company's inventory. An Artificial Neural Network (ANN) was developed to find an accurate forecasting model for the company's inventory. The result showed that there were two significant factors affected demurrage, initial stock, and forecasting error. The statistical test inspired to developed two strategies in optimizing the demurrage cost, involve reducing forecasting error using the ANN model and increasing the initial stock for safety stock. An ANN model with Conjugate Gradient with Powell/Beale Restarts algorithm, two hidden layers, and 25 neurons was found an accurate model with RMSE value 0.186 for inventory's forecasting. For the initial stock strategy, this paper suggested that the company was expected to continue in maximizing shipments and fulfil more orders on time.