Optimization models based on GM (1,1) and seasonal fluctuation for electricity demand forecasting

In this paper, a novel combined approach which combines the first-order one-variable gray differential equation (GM (1, 1)) model derived from gray system theory and seasonal fluctuation from time series method (SFGM (1, 1)) is proposed. This combined model not only takes advantage of the high predictable power of GM (1, 1) model but also the prediction power of time series method. To improve the forecasting accuracy, an adaptive parameter learning mechanism is applied to SFGM (1, 1) model to develop a new model named APL-SFGM (1, 1). As an example, the statistical electricity demand data from 2002 to 2011 sampled from South Australia of Australia are used to validate the effectiveness of the two proposed models. Simulation and graphic results indicated that both of two proposed models achieve better performance than the original GM (1, 1) model. In addition, the APL-SFGM (1, 1) model, which is actually an adaptive adjustment model, obtains a higher forecasting accuracy as compared to the SFGM (1, 1) model.