Using Forecasting Methods to Increase the Accuracy of Container Demand Requirement: An Indonesian Case Study

PT Crieta Logistics ships goods in containers. In carrying out its business, the company often experiences lost sales which are caused by the unavailability of containers so that the orders can be fulfilled. From January to July 2022, PT Crieta Logistics experienced lost sales of 21, 9, 12, 34, 19, 4 and 2 TEUs respectively. Therefore, in this study forecasting the number of requests for containers at PT Crieta Logistics conducted to make decisions related to handling demand for containers in the coming period. This forecasting result served as a basic input to establish policies of container availability, such as whether to reorganize contracts with vendors or others. The forecasting is done using 3 methods, namely linear, quadratic, and exponential regression with historical data of 2 years from July 2020 to July 2022. In determining the most accurate forecasting method, the error value is calculated using the mean absolute percentage error (MAPE) method. In the linear regression method, the MAPE value is 9.5%, while in the quadratic method, the MAPE value is 9%, and in the exponential method, the MAPE value is 9.2%. The quadratic method was chosen as the method suitably used in forecasting the next 12 periods because it has the smallest MAPE value. Forecasting demand using the quadratic method produces 12,631 units of containers for the next 12 periods, from August 2022 - July 2023.