Evaluating the Financial Viability of Vehicle-to-Grid for Autonomous Electric Ride-Hailing Fleets

Flexibility is a key requirement for future electricity markets. While both smart charging and vehicle-to-grid (V2G) technology have emerged as critical elements to fulfill this need, V2G adoption lags behind. Fleets of autonomous electric vehicles (AEVs) in the ride-hailing industry offer new prospects for the adoption of V2G, but an evaluation of the economic viability remains imperative. We propose a deep reinforcement learning algorithm for an operator of a ride-hailing fleet of V2G-enabled AEVs, seeking profit-maximizing strategies for fleet (dis)charging based on passenger demand and real market prices. By empirically evaluating our algorithm using data from ride-hailing operations in Manhattan, we find that it leads to only a marginal increase in profits compared to smart charging, suggesting the need for additional incentives to encourage V2G adoption. Our findings offer insights into the potential of V2G for ride-hailing and highlight the challenges and opportunities for policymakers to consider.