Modeling and Simulation for an Improved Grey Theory Approach in Cars

There are a number of predictive methods available to forecast market changes. Nevertheless, most of these methods require a large amount of historical data and sophisticated input factors to support the forecasting process. To overcome this limitation, grey theory has been developed. The core mathematical basis is the grey differential equation, GM(1,1), which has similar characteristics to the differential and difference equation as well as the exponential function. By using GM(1,1) as a forecasting model, as few as four data points are required to realize a forecast. It can also cope with both indeterminate and incomplete when solving the grey equation, the horizontal is almost always artificially set to 0.5, which does not always guarantee the smallest forecasting error. In this paper an improved GM(1,1) method is proposed. This model continuously adjusts parameters to minimize the variance of the forecast error, our grey predictor therefore becomes to a dynamic forecasting model.