论文部分内容阅读
近年来,以线段交通量观测值作为约束条件来推求OD出行矩阵模型大量出现,这类模型主要目标是保证OD矩阵的推求质量。一般情况下,仅根据交通量不足以确定符合实际的OD矩阵。因此,评价模型的推求精度就变得很重要。本文采用“最大可能相对误差”(Maximum PossibleRelative Error,缩写为MPRE)的概念对OD出行矩阵推求可靠度进行理论研究。这里提出的最大可能相对误差可以表示为简单的非线性规划问题。它可以给出最优拟合的OD矩阵实际相对误差的上限值,因而在分析OD出行矩阵推求可靠度方面非常有用。
In recent years, a large number of OD travel matrix models have been deduced using the observation of line traffic volume as constraints. The main objective of such models is to ensure the quality of OD matrix estimation. In general, OD matrices that match the actual are only based on insufficient traffic. Therefore, the accuracy of the evaluation model becomes very important. In this paper, the concept of “Maximum Possible Relative Error” (abbreviated as MPRE) is used to theoretically study the reliability of OD trip matrix. The maximum possible relative error presented here can be expressed as a simple non-linear programming problem. It gives an upper bound on the actual relative error of the optimally fitted OD matrix and is therefore very useful in analyzing the OD trip matrix for reliability.