Incentive-compatible mechanisms for continuous resource allocation in mobility-as-a-service: Pay-as-You-Go and Pay-as-a-Package
share
Mobility as a Service (MaaS) has recently received a significant attention from researchers, industry stakeholders, and the public sector. The vast majority of existing MaaS paradigms are articulated based on the traditional segmentation of travel modes, e.g. private vehicle, public transportation (bus, metro, light rail) and shared mobility (car/bike/ride-sharing, ride-sourcing). In the context of `Everything-as-a-Service' (XaaS), service providers have evolved from product-based models towards less segmented representations in which resources are priced in a continuous fashion. Yet, such continuous resource allocation formulations are inexistent for MaaS systems. This study attempts to address this gap by introducing innovative MaaS mechanisms that allocate mobility resources to users without any form of travel mode segmentation. We introduce an online auction framework where travelers have the possibility to bid for continuous mobility resources based on their requirements and willingness to pay. We propose two MaaS mechanisms, Pay-as-You-Go (PAYG) and Pay-as-a-Package (PAAP), which allow travelers to either pay for the immediate use of mobility services or to subscribe to mobility service packages for a more protracted usage. Both MaaS mechanisms are based on mixed-integer or linear programming formulations designed to maximize social welfare. We show that the proposed PAYG and PAAP mechanisms are incentive-compatible, develop efficient online primal-dual algorithms to implement the proposed MaaS mechanisms and derive theoretical bounds on the worst-case performance of these algorithms. Moreover, we design a rolling horizon framework to incorporate booking flexibility. Numerical results on extensive problem instances generated from realistic mobility data highlight the benefits of the proposed MaaS mechanisms, and quantify the trade-offs among the proposed approaches.
READ FULL TEXT
