Mobility-as-a-Service (MaaS) seamlessly integrates public transit with mobility-on-demand services to provide users with a one-stop tailored mobility service. Motivated by the practices of the SocialCar project, we study the integration of one-to-many peer-to-peer ridesharing and public transit in the morning commute setting. However, it becomes extremely difficult for the platform to optimise matching, routeing and scheduling for the users over a large network. For practical application, we develop a decision model for the integrated matching problem and propose an iterative distributed optimisation algorithm. The algorithm decomposes the original problem into parallel small subproblems within clusters and iteratively updates clustering as guided by theoretical results; furthermore, an incremental approach is adopted when updating clusters to reduce computational complexity. We conduct extensive computational experiments with real-world data, which demonstrate that our algorithm can quickly generate high-quality solutions for MaaS systems in practice while substantially reducing the users' vehicle miles travelled.
Read full abstract