A novel multi-variable grey forecasting model and its application in forecasting the amount of motor vehicles in Beijing. (November 2016)

• Record Type: Journal Article
• Authors: Zeng, Bo; Luo, Chengming; Liu, Sifeng; Li, Chuan
• Is Part Of: Computers & industrial engineering. Volume 101(2016)
• Page Start: 479
• Page End: 489
• Publication Date: 2016-11
• Subjects: Multi-variable grey forecasting model -- NGM(1, N) model -- Modeling and optimizing -- The amount of motor vehicles in Beijing; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2016.10.009
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: A novel multi-variable grey forecasting model, NGM(1, N ), is proposed. A linear correction item and a grey action quantity item are introduced in GM(1, N ). The NGM(1, N ) has a better structure and performance than those of other grey models. The amount of motor vehicles in Beijing is effectively forecasted using NGM(1, N ). Abstract: The structure defect of the GM(1, N ) model is the major reason for its low simulation and prediction performance. To address this issue, a linear correction item h 1 ( k − 1) and a grey action quantity h 2 are introduced into the GM(1, N ) model to improve its structure in this paper. Specifically, the ' h 1 ( k − 1)' reflects the linear relations between the dependent variable and the independent variables, and the ' h 2 ' shows the data change law of the dependent variable sequence. Based on this, a novel multi-variable grey forecasting model, NGM(1, 1), is proposed. Furthermore, the NGM(1, N ) model's time-response expression and the final restored expression are proved, its initial value is optimized, and a MATLAB program for building the NGM(1, N ) model is developed. Lastly the NGM(1, N ) model is applied to simulate and forecast the amount of Beijing's motor vehicles. The mean relative simulation and prediction percentage errors of the new model are only 0.009% and 1.149%, in comparison with the ones obtained from the traditional GM(1, N ) model and the classical DGM(1, 1) model, which are 4.680%, 10.685% and 4.411%, 11.167% respectively. The findings show that the new model has the best performance, which confirms the effectiveness of the structure improvement.