Traffic Flow Prediction Based on Fractional Seasonal Grey Model

Abstract

Based on the seasonal characteristic of urban road traffic flow data and the principle of new information, a new fractional seasonal GM(1, 1) prediction model is proposed. In the new model, a fractional cycle truncation accumulated generation operator(FCTAGO) was firstly proposed to weaken the stochastic disturbances and the seasonal characteristics of the original sequence, and then the particle swarm optimization(PSO) algorithm was adopted to find the optimal fractional order. Finally, the new model was applied to stimulate a trunk road traffic flow of Nantong,Jiangsu Province. The numerical results show that the average absolute percentage of the new model has a fitting error of 8. 1260% and a prediction error of 7. 6216%, which are much better than those of the seasonal rolling GM(1, 1)model, fractional GM(1, 1) model and seasonal discrete GM(1, 1) model.

Keywords

• fractional cycle truncation accumulated generation operator
• particle swarm optimization algorithm
• traffic flow forecasting
• gm （1 1）model