The Construction Of Gray Multivariable Discrete Time Delay Model Based On The Time-varyinginfluence Of Driving Factors

With the development of social economy and the innovation of science and technology,the complexity of events increases,and the uncertainty system becomes more and more common and complicated.As one of the classical prediction models in the grey system theory,the grey multivariable prediction model is scientific and accurate after more than 30 years of theoretical innovation and practical application,considering the overall influence of driving factors on the model.However,in the prediction of energy demand,electricity consumption and grain output,not only the real-time influence of crowd driving factors should be considered,but also the influence of driving factors' changing trend on the feature sequence should be considered.In order to improve the applicability of the model,this article in view of the driving factors on the characteristics of sequence time-delay and time-varying problem,establish TTDGM(1,N)model,the driving factors as time changes on changes of the character sequence function,further clarify the mechanism of action of driving factors,and explore the parameters of the model calculation and responsive solution,and applied to the whole society of jiangsu province electricity consumption prediction research.In view of the problem that the time-varying characteristics of drivers are not considered when the time-varying dynamic changes of drivers are analyzed in the multivariate gray model,a discrete time-varying gray model(TTDGM(1,N))based on the time-varying influence of drivers is constructed by introducing time-varying control,and the method to solve the parameters is discussed.Combining gaussian distribution and particle swarm optimization,the influence of drivers on the behavior sequence of the system in different lag periods is simulated,so as to make full use of the previous data of drivers.TTDGM(1,N)was used to predict the total social electricity consumption of jiangsu province.Firstly,the total social electricity consumption of jiangsu province and its driving factors from 2005 to 2018 were analyzed to clarify their correlation and time delay.Based on the analysis results,the TTDGM(1,N)model was built to predict the total social electricity consumption of jiangsu province,and its applicability,scientificity and accuracy were analyzed.The results show that the model can describe and predict the operation law of the small sample data system when the driving factors have time-varying delay effect on the behavior sequence.Finally,the TTDGM(1,N)model is used to predict the electricity consumption of the whole society in jiangsu province from 2019 to 2020,and combined with the predicted results,corresponding Suggestions are put forward for the relevant departments of energy and power in jiangsu province from four aspects: optimizing the power supply structure,building a new industrial system,balancing the contradiction between supply and demand,and accelerating the power demand response.