Multimodal choice model for e-mobility scenarios

The paper focuses on the definition, calibration and testing of a simulation model that is able to represent multimodal choice behaviours for electric vehicles. Taking into account the interchange between public transport and electric private mobility, the model estimates the parking demand at the Park & Ride sites equipped with charging stations. The model is based on a data-driven approach, in which mainly Floating Car Data and open data of public transport have derived the explanatory variables. Specifically, a machine learning method (Random Forest) has been used to calibrate and test the model in the real case of the metropolitan area of Rome (Italy). We first perform a stability analysis, letting the parameters of the model vary. We then carry out a sensitivity analysis on the variables that can affect the user propensity to adopt the Park & Ride. Finally, we profile and test an incentive policy to boost the choice of Park & Ride. Results suggest that the model succeeds in simulating Park & Ride by electric vehicles and, therefore, it can be extremely valuable for planning financial support to the multimodal travel choice and forecasting vehicle-to-grid scenarios.