The express/local mode of municipal rail transit provides passengers with multiple alternatives to achieve more efficient and superior travel, in contrast to the conventional all-stop operation mode. However, the various route choices (including direct express trains, direct local trains, or transfers) covering different passenger groups pose a significant challenge to passenger flow assignment. To understand route choice behavior, it is crucial to measure the passenger heterogeneity (variability in individual and trip attributes) in order to propose targeted solutions for operation schemes and service planning. This paper proposes a hybrid model by integrating structural equation modeling and the mixed logit model under express/local mode to estimate the impact of passenger heterogeneity on route choice. An empirical study with revealed preference and stated preference surveys carried out in Shanghai revealed how individual and trip attributes quantitatively impact the sensitivity of factors in route choice. The results show that age and trip purpose are more significant factors. Compared to the control group, the probability of express trains is reduced by 10.22% for the elderly and by 11.36% for non-commuters. Our findings can provide support for more reasonable operation schemes and more targeted services.

与传统的全站运营模式相比，城市轨道交通的快车/慢车模式为乘客提供了多种选择，实现更高效、更优质的出行。然而，覆盖不同客群的多种线路选择（包括直达快车、直达慢车或换乘）对客流分配提出了重大挑战。为了了解乘客的路线选择行为，衡量乘客的异质性（个人和行程属性的变异性）至关重要，以便为运营方案和服务规划提出有针对性的解决方案。本文提出了一种混合模型，通过整合结构方程模型和快车/本地模式下的混合logit模型来估计乘客异质性对路线选择的影响。在上海进行的一项针对显示偏好和陈述偏好调查的实证研究揭示了个人和旅行属性如何定量影响路线选择因素的敏感性。结果表明，年龄和旅行目的是更重要的因素。与对照组相比，老年人乘坐特快列车的概率降低了10.22%，非通勤者乘坐特快列车的概率降低了11.36%。我们的研究结果可以为更合理的运营方案和更有针对性的服务提供支持。