论文部分内容阅读
为了估计交通流量,提出了一个使用先验路段流的贝叶斯网络模型.该模型把路段流量设为OD流量的父节点.在正态分布假设下,模型考虑了总交通流水平、路段流可变性以及交通量守恒的随机扰动.根据先验路段流确定所有变量的先验分布.通过更新一些观测的路段流量,给出后验分布.后验分布的方差往往随着路段流量的逐步更新而不断减小.基于得到的后验分布,给出点预测和相应的概率区间.为消除OD矩阵估计和交通分配之间的不一致,组合了贝叶斯网络和随机用户均衡模型,通过迭代得到均衡解.算例结果验证了提出的贝叶斯网络模型和组合方法的效果.
In order to estimate traffic flow, a Bayesian network model using a priori link flow is proposed, which sets the link flow as the parent node of OD traffic. Under the normal distribution assumption, the model considers the total traffic flow level, Variability, and traffic disturbance. The prior distribution of all variables is determined according to the a priori segment flow, and the posterior distribution is given by updating some of the observed segment flows. The variance of a posteriori distribution tends to increase with the gradual updating of the segment flow And then decrease.According to the obtained posterior distribution, point prediction and the corresponding probability interval are given.In order to eliminate the inconsistency between OD matrix estimation and traffic assignment, a Bayesian network and random user equilibrium model are combined and iteratively obtained The result of the example verifies the effectiveness of the proposed Bayesian network model and the combination method.