摘要
短时间尺度交通流预测是解决城市拥挤的关键。随着预测时间跨度的缩短 ,交通流的规律性越来越不明显 ,传统的预测方法难以凑效 ,文章使用自回归求和滑动平均模型 ARIMA(p,d,q)对短时交通流数据进行建模预报 ,提出以城市交叉路口信号灯时间周期作为动态流量数列的时间刻度。实际应用表明 ,该模型取得良好的预测效果。文章提出的对缺失值的处理方法 ,提高了模型的适用性。
Short-term prediction is the key to solve the congestion in cities. With the s hortening of counting interval, the regularity of the traffic is not clear and t he traditional prediction method does not work. This paper uses ARIMA(p,d,q) m odel to predict the short-term traffic flow. And it uses the cycle of the traff ic lights at the crossroad in cities as the counting interval. As a result, the prediction model work well. The paper also presents a method to deal with the ab sent data, which has improved the applicability of the prediction model.
出处
《交通与计算机》
2005年第1期26-30,共5页
Computer and Communications