随机用户均衡条件下反推OD矩阵的可靠性研究

Reliability of OD Matrix Estimation Based on SUE Model

由于随机用户均衡模型求解的复杂性,人们对基于该模型的反推OD矩阵的可靠性提出质疑。针对这一问题,文章以交通规划软件TransCAD为实验平台,通过所设计的实验对随机用户均衡条件下反推OD矩阵的可靠性进行了研究。实验结果表明:路段流量样本和初始OD矩阵等输入数据对反推OD矩阵可靠性影响明显,增加路段样本量或降低初始OD矩阵的偏差均能有效的提高反推OD矩阵的可靠性。

The reliability of OD matrix estimation is very important to the application of estimation method. Due to complex mathematical formulation and solution of Stochastic User Equilibrium （SUE） model, the reliability of OD matrix estimation based on SUE model is still doubted in practice. In this paper, by taking the software TransCAD as a platform, an experimental research on the reliability of OD matrix estimation based on SUE model was carried out. The experimental result indicates that the reliability of the estimated OD matrices is greatly dependent upon the input data, mainly including link flows and base OD matrices, and that increasing link flow samples or decreasing the discrepancy of base OD matrices can effectively improve the reliability of the estimated OD matrices.