Energy Demand Prediction by Using Neural Network based on Patricle Swarm Optimization

Energy has essential role in the production process and social welfare and its demand prediction is esential for regulate the market and the supply. Due to volatilitys and non-linearity of energy demand and its variables، the non-linear models espicialy neural networks (NN) and paricle swarm optimization (PSO) have been more sucsees in this regard. As respects to their weaknesses such as imposing the specific form، necessity to the larg samples and failur to finding global optimum، in this study these shortcomings fixed by combining thes methods as hybrid algorithm. After applying and comparing this technique with common techniques on energy demand prediction between 1967 -2011، the results confirm higher predictive performance of hybrid technique and the explanatory power of the used variables.