Electricity Load Forecasting Using a Hybrid Artificial Neural Network and PSO Approach

Abstract

Electricity demand forecasting is crucial for effective energy management, particularly in regions with unique consumption patterns like Palangkaraya. This study explores the application of Artificial Neural Networks (ANN) enhanced by Particle Swarm Optimization (PSO) to improve forecasting accuracy. While traditional ANN models offer robust capabilities for handling nonlinear data, their performance is often limited by the challenge of identifying optimal model parameters. By integrating PSO, this research aims to refine parameter selection, thereby enhancing the predictive precision of the ANN model. The study compares the performance of the standalone ANN model and the ANN-PSO hybrid model. Results demonstrate that the ANN-PSO model significantly outperforms the standard ANN, achieving a Mean Absolute Percentage Error (MAPE) reduction from 10.34% to 3.05%, and Root Mean Square Error (RMSE) improvement from 1,061,485.57 to 366,879.94. These findings underscore the capability of the ANN-PSO approach to better capture the intricate patterns in electricity consumption data. The proposed model adapts effectively to the localized energy consumption characteristics in Palangkaraya, offering actionable insights for energy planners. By leveraging the strengths of both ANN and PSO, this research contributes to the development of more accurate and reliable forecasting tools, paving the way for optimized energy management strategies.