City traffic flow prediction based on improved GM(1,1) model

This paper studies the short-term traffic flow prediction method which is based on an improved GM(1,1) model. Considering the defect of the grey action, we change the grey action b into a dynamic variable b(1)sinpk+b(2) which contains the time term because of the traffic flow undulating. Thus, we construct a simple trigonometric grey GM(1,1 vertical bar sin) model, give out its time response equation, analysis the scope of the p value based on the Taylor expansion of the time response equation. By contrasting to the GM(1,1) model, we conclude that the GM(1,1 vertical bar sin) forecasting model has higher prediction accuracy then the traditional GM (1,1) model for undulating data.