| The finite road resource in the city is difficult to load the rapidly growth of traffic flow, so as to result in the traffic congestion, and traffic congestion prediction is one of most important tools for solving the traffic congestion. However, because the influential factors of traffic system is very more, and various traffic parameters contain strong randomness, the congestion prediction is difficult to carry out, the prediction successful rate and reliability is not very higher. In view of this problem, this paper will build the Gray-weighted Markov model to cater to the need of traffic congestion prediction. The detailed process is following points:Firstly, one the basis of literature review, this paper presents the definition, classification, reasons, characteristics of the traffic congestion, and illustrate some classical traffic congestion identification algorithm and velocity prediction model.Then, this paper discuss the relation between the velocity prediction and congestion identification, illustrate the theory of velocity-based congestion prediction model, determines the velocity threshold standard, build the congestion prediction model based on Grey prediction theory and Markov theory, and develop this model by adding the weight for getting a better prediction result.Finally, the model was applied in the case of Jianshe road in Shijiazhuang, to predict the traffic state of6different times in the future4days. The prediction result of this model also was compared with Gray GM(1,1) model and Gray GM(1,1)-Markov model. The results show, the identification successful rate of this model is over than66percent, it is better than Gray GM(1,1) model and Gray GM(1,1)-Markov model. So it proves this model have a good reliability and avail in the traffic congestion prediction. |
