基于路段转向流量的拥挤路网OD矩阵估计

OD matrix estimation based on turning traffic flow for congested networks

传统的OD矩阵估计方法大部分都是基于路段流量的,由于路段流量数目远小于OD对的个数,因而限制了这些方法的推算精度。针对拥堵路网,提出了一种基于路段转向流量的OD估计方法,以提高OD估计的精度。分析了路段转向流量能够降低OD的可行解集的范围。通过双层规划模型求解拥挤路网上OD估计问题。由于最大熵模型不依赖于先验OD矩阵,可以应用到更多的OD估计场景中,因此上层模型采用的是最大熵模型,下层采用用户均衡模型。实验结果表明:基于路段转向流量可以增加估计的精度。

Conventional methods for estimating origin-destination（OD） trip matrices are primarily based on link traffic counts. In a real traffic scenario, due to the numbers of links are much smaller than the number that of OD pairs, it greatly limits the accuracy of these methods. This paper proposed an OD matrix estimation method based on turning traffic flow to improve estimation accuracy in congested networks. This method can considerably reduce the range of feasible solution set. The bi-level programming model is further applied to solve OD matrix estimation problem. The upper model is the entropy maximization（EM）model which is independent of the prior matrix. The lower model is the user equilibrium（UE） model. The experimental results show that our method can increase estimation accuracy effectively.