Prediksi Kebutuhan Stok BBM pada SPBU Pagas Narita Soekarno Hatta Menggunakan Metode ARIMA

This study aims to comprehensively analyze sales patterns and forecast fuel stock requirements at Pagas Narita Soekarno Hatta Fuel Station using Time Series Forecasting with the Autoregressive Integrated Moving Average (ARIMA) approach. Monthly sales data from January 2025 to June 2026 covering Biosolar, Pertalite, Dexlite, and Pertamax were examined in detail. The methodological steps included checking stationarity, applying transformation or differencing, identifying ACF and PACF, estimating parameters, conducting Ljung-Box diagnostic tests, and generating forecasts. Model accuracy was assessed using MAE, MSE, and MAPE to ensure reliability. The selected models were ARIMA(1,0,0) for Biosolar, ARIMA(1,0,1) for Pertalite, ARIMA(0,1,1) for Dexlite, and ARIMA(0,1,1) for Pertamax. The respective MAPE values were 22%, 6%, 57%, and 53%, indicating varying levels of precision. Forecasts for July–December 2026 indicated average monthly requirements of approximately 790,800 liters of Biosolar, 175,400 liters of Pertalite, 8,000 liters of Dexlite, and 21,130 liters of Pertamax. These findings provide valuable insights to support effective fuel inventory planning and strategic decision-making for operational sustainability.