Interactions in mobility-as-a-service eco-systems

Background / motivation Mobility-as-a-Service (MaaS) integrates multimodal and mobility sharing services into a single bundle of mobility services to customers for use in their daily trips. This new mobility concept is regarded as a solution for reducing congestion and pollution levels by reducing private car usage. In future cities, multiple MaaS brokers may co-exist with public and private transport service providers. They will need to offer attractive services to customers (to compete with private car use) and contribute to sustainable policy goals of the government(s) (to justify permissions, subsidies), while at the same time being commercial and aiming at profit. This leads to the emergence of a highly flexible and adaptive eco-system of mobility stakeholders, whose interactions may be dynamically influenced by policy measures (permissions, subsidies, taxes,…), competition, strategic alliances etcetera; and all this is subject to changing preferences in the market and constraints imposed by the physical transportation infrastructures. The goals of these stakeholders are not necessarily aligned, and each stakeholder disposes of only a few decision variables that he controls. However, his satisfaction (e.g. profit) depends on the combination of the variables he controls with variables controlled by others. For instance, travelers can choose the travel mode they use, but the price is determined by the service provider. A shared-taxi provider may choose the tariffs for its service, but is constrained by the willingness to pay of his customers, which depends in turn on the tariffs of the alternative, competing travel services that exist in the market. In addition, he may be taxed or subsidized by a government who wishes to exert some control over his market, e.g. to stimulate him to not compete with mass transit, but rather offer complementary services like feeder trip collection. The shared-taxi dispatch may choose the sequence in which he picks up and drops off passengers and the path between these points, but is confronted with the constraints of the physical infrastructure that may exhibit congestion if too many other road users choose the same path. Research objectives The MaaS eco-system with the interacting stakeholders can be considered as a multi-player game. Mathematical techniques have been developed to model dynamic interactions and their resulting equilibria in similar techno-economical systems like energy production and distribution markets (Delarue et al., 2018; Poncelet, 2018), and supply chain networks (Nagurney et al., 2002). When the optimality or equilibrium conditions of each stakeholder can be analytically described in simple formulas, joint equilibrium in these systems has been modeled as mixed complementarity problems (MCP), which are a special case of variational inequality problems (VIP (Nagurney, 1998)). In fact, many efficient solution algorithms for simpler, more traditional transportation system models like route and departure choice equilibrium models, as well as the theoretical proofs of their convergence and uniqueness, also find their origin in VIP-theory (Friesz et al., 1993; Smith, 1979). Given their related mathematical structure captured by VIP formulations, it should thus be possible to combine traditional transportation network modeling and novel interactions in the MaaS eco-system into a joint modeling framework with corresponding solution algorithms (Nagurney, 2006; Xu et al., 2015). The aim of this research is to develop such models for the MaaS ecosystem, for exploration of equilibria and dynamic interactions in these systems and the factors influencing them. This aim is too broad to be captured in one research and needs further delineation and specification. This will be part of the exploratory phase. On the one hand, the state of the art of similar efforts will be reviewed and open issues identified. On the other hand, the opportunities provided by the research context of the candidate at KU Leuven will be inventoried. For instance, at the department of Mechanical Engineering, there is a long tradition of developing models and algorithms for (dynamic) network traffic and equilibrium modeling (Himpe et al., 2019, 2016; Yperman et al., 2006), vehicle routing (Beullens et al., 2003; Vansteenwegen and Gunawan, 2019), and MCP modeling of energy markets (Delarue et al., 2018). The challenge is to exploit these opportunities for progress in modeling MaaS eco-systems.