Forecasting India’s electricity consumption using particle swarm optimization

Abstract

This paper uses Particle Swarm Optimization (PSO) technique to estimate the electricity consumption in India, based on economic indicators. The data used to estimate the consumption are non-linear. An exponential model is developed and applied to forecast the electricity consumption based on the economic indicators such as population, per capita Gross Domestic Product (GDP), import and export data. The available data are partly used for training the model (1975–2000) and remaining used for testing the model (2000–2010). Mean Absolute Percentage Error (MAPE) is used as an evaluation criterion for finding the future electricity consumption up to the year 2025. The results are compared with the 18th Electric Power Survey of India.