Optimal Alpha Level Setting in GM (1, 1) Model Based on Genetic Algorithm

The grey forecasting model, GM (1, 1), with the property of processing with a minimum of data, has been successfully applied in various fields. However, it has been discovered that different errors may be directly induced by different alpha levels in predicted operations. Accordingly, the parameter α plays an important role on forecasting. Thus, how to search for the optimal setting of parameter α is a valuable work. In this paper, Genetic Algorithm (GA) method has be applied in GM (1, 1) model for handling this problem. We present two illustrative examples to compare between Deng's method and our revised method. These results are useful in that they diminish the margin of error.