Abstract: In order to determine the sample interval that has the highest similarity to the annual highway traffic flow, and provide a reliable sample selection scheme for traffic flow analysis, a similar evaluation index system based on three dimensions of time, space and axis was established based on the oneyear highway truck toll database data in a domestic province. The Euclidean distance, the angle between the plane vector and the entropy weight method are used to calculate the dispersion between each sample and the annual data. The results show that: when considering a certain dimension alone, the results are different and cannot represent the overall characteristics. Under the similar evaluation system, the samples with the lowest dispersion are March, the eighteenth week and Monday.

Key words: traffic engineering, holographic sample, similar evaluation system, highway freight toll flow database, entropy weight method

摘要： 为了确定与全年高速公路交通流量相似度最高的样本区间，为交通流量分析提供可靠的样本选择方案，以国内某省一年高速公路货车收费流水数据库的数据为对象，建立基于时间、空间、轴型三个维度的相似评价指标体系。采用欧氏距离、平面向量夹角公式以及熵权法来计算各样本与全年数据之间的离散度。结果表明：单一考虑某个维度时，结果有差异，无法代表整体特征；在相似评价体系下，离散度最低的样本分别是三月、第十八周和周一。

关键词: 交通工程, 全息样本, 相似评价体系, 高速公路货运收费流水数据库, 熵权法