Energy Consumption Forecasting of Iran Using Recurrent Neural Networks

Abstract

In this paper, a recurrent neural network model is developed in order to forecast the energy consumption as a complex nonlinear function of gross domestic product (GDP) and population in Iran. This intelligent model is trained by total energy consumption data as output and the population and GDP as inputs during 1976–2001, while 5 annual data points of the following years (2002–2006) are used to validate the model. It can describe time dependencies efficiently and the convergence rate is much faster. This model forecasts the trend of energy consumption annually. Simulation results show that this model can predict energy consumption in Iran with acceptable accuracy. It is expected that this study will be helpful in developing highly applicable energy policies.

Keywords:

• artificial neural networks
• energy consumption
• Iran
• recurrent neural networks