Grey Verhulst Extended Model And Its Application In Energy Consumption Prediction

Objective and effective prediction of energy consumption not only optimizes the energy consumption structure,but also provides important information for the government to formulate energy protection and emission control measures.With the significant development of new energy sources and changes in the global consumption structure,historical energy data that are more distant from the present may no longer be reliable for forecasting,resulting in a reduction in the amount of energy information,and the grey forecasting theory,which is applicable to "poor information",has attracted attention.Based on the energy forecasting model,the current research status of grey model and the characteristics of energy data,this thesis selects grey Verhulst as the basic prediction model.In view of the inherent shortcomings of the existing grey Verhulst model,using the modeling mechanism of the Verhulst model and meeting the actual needs of energy consumption,three grey Verhulst extended models are studied and applied to forecast a variety of energy consumption.The main contents of this thesis are as follows:(1)In view of the background value error of Verhulst model,this thesis introduces the extrapolation method to construct a new background value,and uses particle swarm optimization algorithm to find the optimal background value coefficient.Finally,the effectiveness of the extended model is verified by comparing various evaluation indexes with other optimization Verhulst models.(2)In view of the dependence of traditional Verhulst model on saturated S-shaped data,this thesis establishes a prediction model on the Ricaati equation using the relationships among the classical Ricaati equation,the logistic equation and the whitening equation of the Verhulst model,increases the nonlinearity of the Verhulst model's whitening equation,and uses simulated annealing algorithm to find the optimal order of the nonlinear terms.Finally,the validity of the model is proven by a variety of model evaluation metrics and practical cases.(3)In view of the fact that the energy consumption is restricted by other factors and the problem that the univariate Verhulst model may have insufficient fitting ability,the main influencing factors of energy consumption are selected by the grey correlation analysis method,and the energy consumption differential equation is transformed into the grey multivariable Verhulst model through the grey difference information principle,then the extrapolation background value method and genetic algorithm are used to find the most accurate value of the background value in the system.Finally,an actual case is used to illustrate the effectiveness of the multivariable model.(4)After completing the extended study of the grey Verhulst model,the three models obtained are used to forecast six common energy sources,including coal,natural gas,nuclear,hydroelectric,oil,and common new energy sources.The results of six numerical experiments show that the three models can effectively predict the energy consumption.Finally,some suggestions on energy consumption are put forward according to the prediction trend.