主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

Chinese Journal of Management Science ›› 2022, Vol. 30 ›› Issue (3): 221-229.doi: 10.16381/j.cnki.issn1003-207x.2020.2182

• Articles • Previous Articles     Next Articles

Construction of Linear Time-varying Parameters DLDGM(1,N) Model Based on Driving Factors Control

LI Ye, DING Yuan-ping   

  1. School of Information and Management Science, Henan Agricultural University, Zhengzhou 450002, China
  • Received:2020-08-21 Revised:2020-12-15 Online:2022-03-19 Published:2022-03-19
  • Contact: 李晔 E-mail:zzliye@163.com

Abstract: Aiming at the problem that the dynamic change characteristics of parameters with time is not considered, and the mechanism of driving factors is not clarified in the traditional GM(1,N) model, a linear time-varying parameters discrete GM(1,N) model based on driving factors control is constructed by introducing linear time-varying parameters and driving factor control function, which is abbreviated as DLDGM(1,N) model. The new model has more reasonable modeling process, more stable structure, and the traditional multi-variable model’s defects are eliminated in it. To identify the mechanism of driving factor control parameters, calculation methods based on the two situations of sufficient and lack of whitening information are given. In addition, it is proved that DLDGM(1,N) model is entirely compatible with the single variable grey prediction model GM(1,1), DGM(1,1), NDGM(1,1) and TDGM(1,1), and the multi-variable grey prediction model DGM(1,N) and DCDGM(1,N) by adjusting the parameter values,. Finally, to verify the practicality and effectiveness of DLDGM(1,N) model, the model is used to simulate and predict the grain yield in Henan Province and the sample data is obtained by Henan Statistical Yearbook. The mean relative simulation and prediction percentage errors of DLDGM(1,N) model are 0.49% and 1.23%, in comparison with those of DGM(1,N) model and traditional GM(1,N) model, which are 1.00%, 3.12% and 4.68%, 5.25% respectively. The results show that DLDGM(1,N) model has the best performance in simulation and prediction, which on one hand testifies the effectiveness of improving model’s structure, and on the other hand proves that the model has good practicability in the system with time-varying parameters and complex driving factors mechanism.

Key words: grey prediction; linear time-varying parameters; GM(1,N) model; driving factors; grain yield forecasting

CLC Number: