基于傅里叶级数残差修正TDGM(1,1)的车流量预测模型

Traffic Flow Forecast Based on Fourier Series Residual Modification TDGM(1,1) Model

车流量预测是城市智能交通研究中的热点和难点问题之一。然而,车流量受到诸多因素的不同程度的影响,使用单一模型难以对其进行准确预测。针对这一问题,本文提出了基于傅里叶级数残差修正TDGM(1,1)的车流量预测模型。该模型首先应用离散灰色模型TDGM(1,1)对原始车流量序列进行建模,并得到初始预测值以及残差序列;然后通过傅里叶级数对残差序列进行二次拟合,同时对预测结果进行修正。通过实例分析以及对比试验表明,该模型可以有效提高车流量预测精度。

Traffic flow prediction is a key problem in urban transport system.However,many complicated factors have impact on the traffic flow prediction,which means that single model can not be used to forecast the traffic flow correctly.Aiming at this issue,a combined residual modification TDGM(1,1)model based on Fourier series is proposed for predicting the traffic flow.Firstly,this model used the TDGM(1,1)model to predict the original series of the traffic flow,and obtain the initial predicted values and the corresponding residuals.Then the Fourier series were introduced to modify the residual series and get the final predicted values.The experiments demonstrated that,in comparison with the common methods,the residual modification TDGM(1,1)model based on Fourier series can improve the prediction accuracy effectively.