Autonomous Mobility-On-Demand (AMoD) provides new options for the morning commute problem. The flexibility of AMoD could help to boost the attractiveness and accessibility of Public Transportation (PT). Intermodal AMoD systems could become a competitive alternative to personal cars. However, considering the convenience, comfort, and expected low fares of autonomous vehicles, there is a risk of competition between privately operated AMoD and PT. The joint design of PT and AMoD can foster their cooperation. This study investigates the joint PT-AMoD design problem in a many-to-one multimodal corridor where three transportation alternatives are available: full personal car on a congested freeway, walking and massive rapid transit (MRT), or autonomous vehicle (AV) and MRT. We introduce a simple dynamic model incorporating a time-dependent mode and route choice subject to user equilibrium (UE) constraints. The model presented: (i) accounts for how UE settles and evolves, (ii) provides insight on PT-AMoD cooperation opportunities and competition risks depending on the design choices, (iii) is compatible with design optimization heuristics. We apply the model to a realistic scenario based in the city of Lyon (France). The number of MRT stations, their locations, the number of AV fleets, and their coverage zone boundaries are optimized with a metaheuristic. The optimization is conducted according to three policies regarding AMoD (protectionism, opportunism, liberalism) and three priority objectives (maximize MRT usage, minimize travel times, reduce car usage). We formulate recommendations for the transportation authority by evaluating the potential benefits of each policy.
