Poster 29: A Multi-Agent Simulation Model of Door-2-Door Vs. Stop-2-Stop Shared-Automated-Vehicle Mobility Services
(Room Palms Ballroom)
16 Jul 19
5:30 PM
-
7:00 PM
Tracks:
Transportation Planning Impact Assessment
In late 2018, “Waymo” launched a self-driving taxi service in Arizona, USA. In the near future shared automated vehicle (SAV) services could dominate the urban transportation scenery (Litman, 2017) and shared by travelers like ridesharing services nowadays (Fleischer et al., 2018). With the advancement of information and communication technologies and proper regulation, SAVs may become public domain potentially improving air quality, road safety, etc. (Fagnant & Kockelman, 2014). Future SAV services could choose to operate on a “Door-2-Door” (D2D) basis – akin to “Uber-Pool” and ”Lyft” picking up customers at their point of origin and dropping them off at their destinations or on a “Stop-2-Stop” scheme (S2S) similar to “VIA” where travelers are picked-up/dropped-off at the nearest transit stop. To estimate these schemes' benefits, we employ a large-scale spatially-explicit multi-agent-based simulation of a future SAV service in the Tel-Aviv Metropolitan area (TAMA), Israel. The simulations at the spatial resolution of individual travelers are performed with the MATSim framework (Horni et al., 2016). Based on the likely results of this study, A 50K fleet of 4-seat SAV vehicles is necessary for performing all internal trips in TAMA. However, long waiting time will likely de-motivate travelers to switch mode. To reach acceptable Level-Of-Service, the D2D fleet should be doubled to 100K, while S2S should be increased to 75K. About half of the vehicles in both schemes will remain vacant even during peak periods. The operational implications of D2D and S2S schemes are further discussed.