Evaluation of Dynamic Public Transport Pricing with MATSIM

Abstract

With emergence of digital fare payment systems, the resolution of problems such as growing amount of air pollution or concentration of congestion around peak hour be- comes feasible by the implementation of dynamic pricing schemes. But, because of the complexity and the difficult acceptance of these strategies by the populations, re- search need to be accomplished beforehand in order to show clearly the benefits brought. In this thesis, the agent based simulator MATSim is used to study effects of First-Best pricing strategy and to develop feasible Second-Best scheme in the case of a rather simple corridor scenario. In addition to that, effects of crowdedness in public transport vehicles and friction effects at boarding are implemented in MATSim and their consequences on the trip distribution of evaluated. Firstly, the implementation of the new dwell time model shows that the dwell time may be strongly increased and the trip distribution spread if boarding occurs in a ra- ther to high filled bus. Secondly, the addition of crowding effects leads up to results such as peak spreading, diminution of the travel times at peak and increasing of the percentage of PT users. The last point even suggests that the implementation of a crowding effect may ameliorate the congestion state of the roads. Thirdly, the imple- mentation of externalities increases the benefits cited above and reproduces real be- haviors: PT users with higher flexibility make changes in their plans in order to avoid the periods where the fares are higher. Finally, a Second-Best pricing strategy using the opportunities given by Smart-Card based fare payment systems has been de- signed. These results confirm that MATSim may be appropriate in order to develop pricing strategies. In further steps, tests of the new implementations in more realistic networks and comparison with real situations shall be done.