Hani Mahmassani - Northwestern University As autonomous vehicle technologies continue to improve, deployment in actual urban networks offers many scenarios for the supply of shared fleet mobility services. These range from services that resemble existing Transportation Network Company (TNC) offerings but without drivers, to services that resemble short-term rentals, and services and business models that do not yet exist but can be imagined by entrepreneurial companies. Such services may be operated with varying degrees of centralization in terms of fleet assignment, scheduling and routing, to achieve either company objectives or societal objectives or a combination of both. Such services may compete with existing urban transit services, or may be integrated with those as part of a Mobility as a Service (MaaS) portfolio-- or somewhere in between. Similarly in the freight and logistics arena, introduction of autonomous vehicles at various steps of the supply chain, from long-haul to last-mile delivery, opens potentially significant opportunities to reduce overall logistics costs along with negative externalities. It also raises operational challenges for interlining autonomous and human-operated portions of the supply network. Deploying autonomous fleet mobility services opens many service opportunities, and gives rise to new classes of operational problems that call for dynamic optimization of many-particle systems on a large scale. We provide several examples of these problems, and approaches that have been developed to address them. We also discuss integration of autonomous fleet mobility services with existing public transit so as to maximize social mobility with a given level of resources.