LAUSR

Abstract Shared autonomous electric vehicles (SAEVs) are expected to serve a significant fraction of the passenger transportation needs in cities and surrounding urban areas. In this paper, we consider optimal operation of a cyber-physical system (CPS) comprising a large fleet of SAEVs and a set of charging hubs located across the transportation network and supported by the power grid. The hubs are considered to have a number of charging stations, a stand-alone battery bank for energy storage, and limited rooftop photo-voltaic (PV) generation capacity. We developed a robust mixed integer linear programming model. It considers a number of practical features of both the power and transportation systems, including day-ahead load commitment for electricity via an alternative current power flow model, real time price spikes of electricity, energy arbitrage, uncertainty in passenger demand, and balking of passengers while waiting for a ride. We demonstrated our methodology by implementing it on a sample CPS with 500 SAEVs and five hubs with fifty charging stations in each. Our methodology yields operational decisions for day ahead commitment of power and real time control of the SAEVs and the hubs. The sample CPS is used to examine impact of hub capacity and fleet size on various system performance measures. We discuss the computational challenges of our methodology and propose a simplified myopic approach that is capable of dealing with much larger fleet sizes and a variety of hub capacities. Reduction in computation time and the optimality gap for the myopic approach are examined.