Effect of Congestion Pricing on Users' Mode Split Using Stated Preference Technique

Congestion pricing is one of the main strategies for demand management in urban areas. One of the main questions for decision makers, for implementing this strategy, is the effect of congestion pricing on users' mode split. For defining this, the willingness to pay of users should be estimated. In another word, we should know that how much users still will to pay for using their personal cars and how they intend to perform their mobility according to various prices. Previous studies had pay less attention to paratransit mode according to pricing. In this research, the effect of congestion pricing on users' mode choice has been investigated. The restricted traffic zone of Tehran has been selected for case study. Due to impossibility of direct observation, the stated preference method was applied for data gathering. Various pricing scenarios, based on experimental design concept, were defined and several types of questionnaires were designed. In these questionnaires, 3 types of data were asked from interviewees including trip chain of the users, socio economic characteristics and pricing scenarios. According to pricing scenarios, 4 choices including using personal car, public transportation, taxi and cancelling the trip were presented to users which they should choose only one option due to every pricing scenario. The reliability of questionnaires have been investigated with cronbach's alpha which results showed the proper reliability of questionnaires. More than 3500 interviews were performed and after preliminary validation, were entered in the database. Based on this data, more than 70 variables were defined which their correlation was estimated and proper variables were chosen. For modeling, due to discrete nature of data, multinomial logit model was applied and calibrated for every choice. In this regard, the feasibility of applying nested logit model was also tested which results showed the invalidity of this model. More than 200 models have been calibrated and finally best validated models have been chosen for describing the mode choice of every alternative.
Results of modeling showed that having more expansive cars will increase the utility of using personal cars and reduce the public transport utilization. Also the residence location of the users is effective in their mode choice. Living in restricted zone, increase the willingness of users to use transit and paratransit mode. Users' education is also important in their choice. People with higher education level have more willingness to pay. The sensitivity analysis showed that pricing can be a proper tool for managing personal vehicles demand. When entrance toll is equal to 3000 tomans, only 35 percents use their personal vehicles. In this condition, 30 percents of users choose transit mode. Results also showed that choosing transit and paratransit mode is similar. When the toll is more than 17000 tomans, the rate of using taxi have higher growth comparing to transit utilization. Also, results showed that the demand for entering to restricted zone can be assumed inelastic. In highest toll, only 10 percents of users cancelled their trip. Finally, the sensitivity analysis for every mode has been accomplished and multiple future researches have been proposed.
Keywords: Mode Split models, Congestion pricing, stated preference method