Comparative Analysis of Grey GM(1,1) and Grey Verhulst Models for Forecasting Electricity Consumption in Indonesia

Abstract

Electricity energy demand in Indonesia continues to increase in line with economic development, technological advancement, and population growth. This condition necessitates appropriate energy planning to ensure a stable electricity supply. Electricity consumption forecasting is required to predict future demand and to serve as a basis for formulating sustainable energy policies. However, studies comparing different grey forecasting models for Indonesia’s electricity consumption remain limited, particularly in identifying the most suitable model for its growth characteristics. This study aims to compare the performance of two models within Grey System Theory, namely the Grey Model GM(1,1) and the Grey Verhulst model, in forecasting electricity consumption in Indonesia. The data used consist of annual electricity consumption data for the period 2000–2024, obtained from the official websites of Statistics Indonesia (BPS) and the Ministry of Energy and Mineral Resources (ESDM). The analysis stages include model construction, parameter estimation, forecasting, and accuracy evaluation using the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). The results indicate that the Grey Verhulst model outperforms the GM(1,1) model, with an RMSE of 10,955.97, an MAE of 9,233.64, and a MAPE of 5%, whereas the GM(1,1) model yields an RMSE of 23,562.39, an MAE of 15,151.57, and a MAPE of 9%. These results suggest that the Grey Verhulst model provides a better fit for the observed data, which exhibits nonlinear growth behavior with a tendency toward saturation. The best-performing model, namely the Grey Verhulst model, produces a forecast of electricity consumption for the year 2025 of 388,025 GWh. This result is expected to serve as a reference for national electricity policy planning.