Investigation and prediction of traffic flow in holidays in Zhejiang section of Shenhai freeway

This paper investigates the traffic data in holidays within Zhejiang province in Shenhai freeway from the spatial and temporal aspects. The results shows that Tangxia monitoring site has a much higher flow than other sites which indicates a heavy traffic jam in holidays. A further Investigation of traffic flow in Tangxia site shows a common higher traffic flow in the north direction compared with that in the south direction. The temporal distribution of traffic flow in each holiday has its own characteristics which are closely related to the characteristics of the holidays. The hourly distribution of traffic flow in each day has the similar trend with high value at daytime and lower flow at night. The artificial neural networks (ANN) model was used to predict the traffic flow in 2014 which was trained with data in the same site and holiday in 2013. The multi-layer feed-forward perceptron networks (MLP) with three-layer structure was applied to predict the traffic flow in the next 5 minutes with the past 25 minutes flow data. An example of Yandang monitoring site shows a pretty well prediction results with mean absolute relative error (MARE) and max absolute relative error (MAXARE) being 0.0214 and 0.1244, respectively. The investigation and prediction of traffic flow can provide a theoretical basis for the transportation administration to make decisions.