Research On Development Of Discrete Grey Model And Its Application

This article to discrete grey model and develop modeling mechanism, characteristics andapplication of the model of a main line. To approximate homogeneous index series, non-homogeneousindex series, a sequence of logarithms, approximately S type sequence, a sequence of oscillation aswell as discrete grey model based on stochastic oscillatory sequence as the main research object.Chinese carbon emissions projections for the actual background, data show different characteristics as astarting point, combined with data transformation technology to deal with the various characteristicsdata make it suitable for discrete grey model. Specific elements include the following:(1) For logarithmic characteristic sequence of grey modeling issues, based on four conditions(improving the smooth level, compression ratio, maintain a concavity, do not increase the restore error)for improving accuracy of grey prediction model, this article constructed a step ratio exponentialtransformation to improve model accuracy:f (x (k)) n kminx (k). The article has proven that the stepratio exponential transformation can meet higher precision grey prediction model of four conditionsand has established based on the logarithm of a step ratio compression logarithmic characteristicsequence of DGM(1,1) model. By examples verify the availability of this transformation, the articleimproved the model’s simulation accuracy, expanded the scope of application of discrete grey model.(2) S type feature sequence of grey modeling problem, according to its curve characteristic, thisarticle will divide it into three phases: Adjustment Stage, Growth Stage, and Saturation Stage.Combined with existing research design of the segmented recursive revision in the dynamicprogramming, the establishment of a segmented recursive revision DGM (1,1) model based on S-shaped feature. And the usefulness of the model was verified by example, which displays theapplication of amended subparagraph to S type character sequence to establish discrete grey modelsimulation with high accuracy.(3) For the grey modeling problem of stochastic oscillatory sequence, by analyzing study on greymodel of an oscillating sequence, this paper found that in the literature most of the accuracy of theirmodels is not stable, for volatile oscillatory sequence simulation is very poor. So this article combinedwith existing research, made up with a positive translation transform and variable weighted geometricaverage transform, through changes in weights of each point in a data series, which can graduallyincrease the smoothness of the sequence and make it’s grey exponential law clear, so as to improve theaccuracy of model simulations. (4) Prediction modeling for China’s carbon emissions problem, this paper has analyzed the1995-2010general characteristics and trends in carbon emissions in China. According to existing researchfoundation and discrete grey forecast model system established based on the data features, this paperhas respectively modeled on effects carbon emissions factors, including population, energy strength,and per capita GDP. Eventually established my carbon emissions volume of DGM (1,4) forecast model,and according to forecast results made has policy recommends.