Study The Mode Choice Behavior In Chinese Bus Rapid Transit Corridors

The rapid development of Bus Rapid Transit (BRT) in China has led to the world’slargest deployed BRT network with the most diversified types. The BRT systems in China’srapid development will continue in the future.As to the gap of there is no study of the mode choice behavior in Chinese BRT corridors,we will be the first to study the mode choice behavior in Chinese BRT corridors. Sixrepresentative corridors in five cities were selected to study, including Beijing BRT1, BeijingBRT3, Changzhou BRT1, Jinan BRT1, Kunming BRT1adn Guanghou zhongshan avenueBRT corridor. Multinomial Logit（MNL）Model, Mixed Logit （ML）Model, Box-CoxMultinomial Logit（BCL）Model, Box-Cox Mixed Logit （BML）Model were built to modelthe choice behavior among BRT, autos, normal buses, and non-motorized. Comparing thelikelihood ratio and hit rate of four models, we found that, except in Guangzhou BRT corridor,BML models were the best. For Guangzhou ML model and BML model, the functiondistributions of random coefficients need to be further study.The traveler’s socioeconomic attributes (gender, age, income) and trip-related attributes(trip purpose, travel time, trip costs, and trip distances) are found to be statistically significantin influencing traveler’s mode choice. Moreover, these characteristics vary significantly bycity and corridor, suggesting BRT deployment strategies must recognize the heterogeneousnature of ridership and travel preferences.In addition, only3-8%BRT users shift from auto users. By sensitivity analysis andscenario analysis, we found saving BRT travel time for BRT users can not significantly attractauto users shifting to BRT. Additional economic and policy constraint for auto users maycause significantly mode shift to public transport.