A nested logit model of transit mode choice of multimodal public transportation using smart card data of Seoul

초록

This study analyzed the transit mode choice in Seoul using data with respect to distance-based fares collected from the City’s Automatic Fare Collection (AFC) system. The distance-based fare system allows a maximum of four transfers with no additional charge to increase transit ridership. In order to analyze the transit transfers, we developed transit mode choice models by defining three utility functions for multinomial logit and nested logit models. The results show that the frequency of transfer is a crucial factor of transit mode choice during peak hours (i.e., 7 a.m. to 9 a.m.). Further, a nested logit structure with an upper level consisting of a single mode (no transfer) and multi-transit mode (with transfer) better explain current transit mode choice behavior. These findings will help international transit agencies and planners who use AFC systems and fare schemes for promoting transit-friendly services and addressing mobility problems.