The objective of this paper was to investigate the factors that influence the travel mode choices of travelers in the Stockholm area. First, a data set including 21,579 original questionnaire results were collected and processed. Then, by referencing result of T-test of variables, two multinomial logit models and two mixed logit models were established respectively. By using Biogeme and Python, all parameters of 4 models were estimated and formal tests of model specification (likelihood ratio) were conducted as well. Finally, a multinomial logit model with log likelihood of -8602.17 was adopted as ultimate model, and it was used for VOT estimation and policy analysis. The results of policy analysis show that policy of decreasing the public transport charges can increase people's willing of choosing public transport but may also increase pressure on public transportation; policy of increasing parking charge performed a relatively low effect; policy of adding the congestion charging for car travel, can reduce people's choice of car travel in a targeted way, while increasing the choice of the other three travel modes in a relatively balanced way. Hence, this study will have positive impact on the transportation decision and policy making in Stockholm region.
Estimating travel mode choices of residents in Stockholm region based on multinomial logit model and mixed logit model
International Conference on Smart Transportation and City Engineering (STCE 2024) ; 2024 ; Chongqing, China
Proc. SPIE ; 13575
2025-04-28
Conference paper
Electronic Resource
English
Appendix F: Multinomial Logit Model
Wiley | 2012
|Transportation Research Record | 2009
|Wiley | 1993
|Multinomial Logit model based on Gumbel Copula
British Library Online Contents | 2009
|Transportation Research Record | 2018
|