交通流的季节ARIMA模型与预报

Modeling and Prediction of Freeway Traffic Flow Using Seasonal ARIMA Models

使用季节ARIMA模型对交通流进行建模及预报为高速公路交通状况分析、道路设施建设、公路效益评价及控制策略设计等提供了一种可靠的方法和途径.介绍了具有周期的季节ARIMA模型的一般表达方式,并提供了使用这一模型进行建模和预报的一般过程.最后以某高速公路的实测数据为例,进行实证分析,得到了72步的长期预报结果,其相对误差为0.

Traffic flow is one of the parameters that affect the studies on analysis of traffic status, traffic facility construction and control structure of freeway. It has obvious seasonal characteristic, therefore seasonal ARIMA models can be a reliable approach to the prediction of traffic flow. A general expression of seasonal ARIMA models with one periodicity is given, and procedures to model and predict traffic flow using seasonal ARIMA models are provided. Feasibility-study experiments show that seasonal ARIMA models can be used to model and predict actual freeway traffic flow. The result of 72-step long-term forecast is obtained and the maximum relative error is 0.13.