According to the U.S. Department of Energy (DOE), the average vehicle occupancy in the U.S. was as low as 1.5 in 2013, contributing to traffic congestion and pollution in many urban areas. The advent of pooled ridesharing services provides an opportunity to mitigate the impacts of low occupancy rates. If widely used, these services could increase the efficient use of transportation capacity, and reduce traffic congestion and emissions. However, some studies suggest that these services have captured public transit demand with negative impacts on the system in general, and transit providers, in particular.
Many urban and rural regions in the U.S. are trying to determine the feasibility and benefits of bridging the first and last mile transit access gap through partnerships between transit agencies and shared mobility providers. The goal is to take advantage of the flexibility of these services and transit benefits. This study evaluated such system. Specifically, the study focused on the potential demand shifts from drive alone mode to this transit-ridesharing program, where individuals walk to a mutual pick-up and drop-off (PUDO) point and pool the ride to the closest transit station. The study developed a novel simulation and optimization framework, and implemented a case study in the San Francisco Bay Area. The framework has three main components. 1) A macro-simulation (activity-based modeling) of the travel decisions which approximates the potential demand for the program. 2) A continuous location-allocation optimization model to optimally locate the PUDOs and allocate demands. And, 3) An agent-based analysis to explicitly simulate the movements of individuals.
The empirical results showed a potential increase of almost 6,000 new daily BART work-related trips in the AM period, from which about 17% switched from driving their own vehicles. Throughout the Bay Area, the system’s users would experience an average waiting time of 7 minutes, a 20-minute in-vehicle time, and had to walk about ¾ of a mile. Around 60% of the new users, could have a combined waiting and access time between 5 and 20 minutes. These are promising results for system efficiency, though they require a reliable transit system for users’ acceptability. To attract more users, and compensate for delays for some, subsidies may be needed.