Optimization Analysis And Application Research Of Grey Forecast Model

Grey prediction model is one of the main research contents of grey system theory,and multi-variable MGM(1,m)model is the extended model of single-variable GM(1,1)model.The current grey prediction model has the following problems:(1)In the process of constructing MGM(1,m)model,the prerequisite selection of the existing background value calculation method is unreasonable.(2)The prediction effect is not ideal when the existing MGM(1,m)model is used to predict the oscillatory original data sequence directly.(3)The current grey prediction model is limited to quantitative forecast,and the quantitative prediction results affected by external factors will often appear large error.In order to solve the above problems,the background value optimization method of MGM(1,m)model,the model and its application based on the optimized background value,and time-varying cloud model are mainly studied in this paper.The main research work is as follows:1.The background value calculation method of MGM(1,m)model is optimized.Existing optimized background value has the problem that the prerequisite selection in the process of constructing background value is limited to the first data point,and the prediction results are affected seriously.Thus,the existing calculation method of the background value is optimized in this paper.Firstly,according to the non-homogeneous exponential function characteristic of the first-order accumulated generation sequence,the non-homogeneous index function is used to fit the first-order accumulated generation sequence.Secondly,the prerequisite is optimized to eliminate the error of the first data point in the original data matrix as the prerequisite.Finally,the calculation method of the background value is reconstructed.Evidence of theory shows that the optimized background value has more reasonable and reliable theoretical basis.2.The OMGM(1,m)model based on monotone sequence and the OOMGM(1,m)model based on oscillatory sequence are constructed.The simulation and prediction accuracy of the existing MGM(1,m)model is unsatisfactory when the original data sequence shows an oscillation characteristic.Therefore,OMGM(1,m)model and OOMGM(1,m)model are constructed based on the optimized background value,the MGM(1,m)model theory is perfected,and the MGM(1,m)model is extended.The experimental results of OMGM(1,m)model and OOMGM(1,m)model show that the monotone sequence and oscillatory sequence can be accurately predicted respectively.In addition,the OMGM(1,m)model is used to solve the prediction problem of foundation pit engineering,which has higher simulation and prediction accuracy,and can offer more accurate deformation condition of foundation pit for the subway foundation pit engineering.3.The time-varying cloud model is constructed.The current grey prediction model is mostly applied in the quantitative prediction.For the defects of quantitative prediction and the limitations of grey prediction model,the time-varying cloud model is constructed.The prediction model expands the application scope of the grey prediction model,and implements effective prediction of the time qualitative concept set.Finally,the correctness and effectiveness of the model is verified by the experimental analysis.