To solve the ill-conditioned problem and the modeling precision in grey prediction control GM(1,1) model, a new simple and effective method is presented. This method is based on multiple transformation to original series and center parallel moving transformation to AGO series, that is, a new grey optimization model based on affine transformation is proposed. Firstly, the parameter relationships and properties of the new model are discussed. Secondly, Based on the condition number and the modeling precision, a multi-objective optimization model is set up. By solving the model and choosing coefficients M and p, the condition number of coefficient matrix can be controlled, meanwhile, the prediction accuracy of the original grey model is improved. Finally, we use an application example for our case study to test the efficiency and accuracy of the proposed method.
[1]
CHENXiao-chuan,et al.
Application of Grey System Theory in Design For Cost （DFC）
,
2003
.
[2]
Xin-Ping Xiao,et al.
Research on ill-conditioned problem in grey prediction control model
,
2004,
Proceedings of 2004 International Conference on Machine Learning and Cybernetics (IEEE Cat. No.04EX826).
[3]
Xiao Xinping,et al.
A new modified GM (1,1) model: Grey optimization model
,
2001
.
[4]
Shiuh-Jer Huang,et al.
Control of an inverted pendulum using grey prediction model
,
1994,
Proceedings of 1994 IEEE Industry Applications Society Annual Meeting.
[5]
Jer Min Jou,et al.
A low-cost Gray prediction search chip for motion estimation
,
2002
.