用多层次分析法从出行者个体、家庭和城市区域3个层面解析了通勤出行时间选择行为的差异性,描述了通勤时间选择的可观测异质性和不可观测异质性,及其相互关系。研究表明:多层线性模型为准确描述出行时间选择行为数据不同层次解释变量和随机项的相互关系提供了有效的分析工具,对交通需求预测和管理工作有重要意义。%T his paper studies the variation of commute departure time using multilevel modeling meth-od .The effects of the nested data hierarchy of person ,household and space on travel choice behavior is explored .By introducing random effects at the three levels and by using explanatory variables with random coefficients ,observed and unobserved heterogeneity in commute departure time choice are ex-amined .The total variation in departure time is decomposed into three components ,namely the inter-personal variation ,the inter-household variation and the spatial variation .T hen based on the validated variation pattern a full model is established to examine how much of the variations can be captured by explanatory variable at different hierarchies .The model is estimated using travel survey data .And the results demonstrate that commute departure time choice is significantly different at the individual , household and space level ,and can be explained partly by the socio-demographic and the commute ac-tivity attributes .It is demonstrated that multilevel modeling method is a useful analysis tool for trav-el ,which meets the needs of simultaneously dealing with unobserved macro-level variations required by the data hierarchy .
展开▼