学术文献库

Mobility-On-Demand (MoD) services have been transforming the urban mobility ecosystem. However, they raise a lot of concerns for their impact on congestion, Vehicle Miles Travelled (VMT), and competition with transit. There are also questions about their long-term survival because of inherent inefficiencies in their operations. Considering the popularity of the MoD services, increasing ride-pooling is an important means to address these concerns. Shareability depends not only on riders attitudes and preferences but also on operating models deployed by providers. The paper introduces an advance requests operating model for ride pooling where users may request rides at least H minutes in advance of their desired departure times. A platform with efficient algorithms for request matching, vehicle routing, rebalancing, and flexible user preferences is developed. A large-scale transportation network company dataset is used to evaluate the performance of the advance requests system relative to current practices. The impacts of various design aspects of the system (advance requests horizon, vehicle capacity) on its performance are investigated. The sensitivity of the results to user preferences in terms of the level of service (time to be served and excess trip time), willingness to share and place requests in advance, and traffic conditions are explored. The results suggest that significant benefits in terms of sustainability, level of service, and fleet utilization can be realized when advance requests are along with an increased willingness to share. Furthermore, even near-on-demand (relative short advance planning horizons) operations can offer many benefits for all stakeholders involved (passengers, operators, and cities).