公路接入段日交通量预测技术优化与应用研究

Optimization and Application of Daily Volume Forecasting Methods for Freeway Extension Sections

为更准确掌握公路接入段日交通量变化趋势,提高扰动路段交通量预测精度,利用曲线拟合法,建立观测路段基年交通量时间序列模型,结合ACF、PACF图例,应用ARIMA技术和Logistic回归曲线拟合方法,进行模型的识别、细化,而后进行参数估计和模型诊断,确定最佳交通量预测模型,完成模型优化.实例应用结果表明,利用ARI-MA技术和Logistic回归曲线拟合方法优化的模型比不考虑观测交通量自相关内容的模型预测精度高,预测误差均值仅为1.53%.

To estimate the trends of traffic volume of freeway expansion more accurately, a time series model was developed based on the time-series curve fitting of highway network traffic. The identified and detailed time-series model was also formulated according to highway network traffic impact factor combined with ACF and PACF Legend. Then, the paper estimated the parameters and diagnosised the model by the Logistic regression and ARIMA technology. The results indicated that the proposed model with ARIMA and Logistic regression was more accurate than traditional models, and the average error is 1.53%.