摘要
对应于城市道路短期交通流复杂的非线性特征,采用一种非线性的时间序列模型来对其变化规律进行探索,以期获得城市道路短期交通流的精确预测。根据现实情况,可以将城市道路的交通流条件划分为两种状态:交通拥堵和交通畅通,在不同的状态下,交通流表现出不同的变化特征,一个二制度自我激励阈值自回归(SETAR)模型的结构能够很好地与之相符。以现实中的城市道路短期交通流数据为样本所进行的实例分析结果表明,被估计模型获得了很好的仿真结果,并能够合理地解释城市道路短期交通流的非线性特征。以此为基础,用估计所确定模型进行城市道路短期交通流的样本外预测,结果表明该模型不仅有较高的预测精度,且预测表现明显优于自回归求和移动平均(ARIMA)模型。
The data of urban short-term traffic flow is characterized by complicated nonlinearity, thus a nonlinear time series model is proposed and employed to explore how traffic flow varies with time so that accurate predictions may be obtained. Traffic flow conditions on urban roads are categorized into two states, namely the congested traffic state and the free-flow traffic state, and traffic flow varies in different ways under different states, which is consistent with the structure of a two-regime self-exciting threshold autoregressive model （SETAR） . The estimates obtained from a case study show the SETAR model can not only offer accurate simulations, but also very well explain the nonlinear characteristics of traffic flow. The out-of-sample prediction performance of the proposed model is compared with the autoregressive integrated moving average model （ARIMA） , and the results show the former does better than the latter.
出处
《土木工程学报》
EI
CSCD
北大核心
2008年第1期104-109,共6页
China Civil Engineering Journal
基金
国家自然科学基金(50608010)
