The Study Of Energy Demand Forecasting Based On The Grey-neural Network Theory

Energy is an important material resource of existence, economy development, social progress and modern civilization. Along with rapid development of economy and society, the energy demand grows continuously. Therefore energy demand forecasting has important theoretic and realistic significance. This paper adopts the methods of systematic analysis, grey system theory and artificial intelligence theory to construct China's energy demand model.Firstly, qualitative and quantitative analyses are applied to analyze the relationship between the energy demand and the influencing factors. Co-integration and Granger causality test are applied to the time series of the total energy consumption and GDP during 1978~2005. We find there is co-integration relationship between China's energy consumption and economic growth. And then the multiple statistical analyses are applied to estimate the relationship among the factors and their impact on energy demand.Secondly, this paper adopts an improved series grey neural network model to forecast the energy demand. According to the advantages and disadvantages of the artificial neural network and grey theory, the forecasting results of grey differential equation are inputted to the neural network with the main influencing factor of energy demand. Thus the new model not only considers the non-linear relationship between the influencing factors and energy demand, moreover it can study the long-term forecasting ability of grey model.Thirdly, we use the data from 1978 to 2005 to build and test model, and then we find that average relative error of forecasting results of this improved series grey neural network model is 1.19%, which is 1.15% smaller than the parallel combined model, 1.08% smaller than the traditional series combined model, the results of this improved series grey neural network model which applied to forecast the energy demand in 2010 and 2020 show that the total energy demand will be 2.4994 billion ton standard coal and 4.0184 billion ton standard coal, which have the reference value for making the energy policy.