Abstract
Fleets of electric vehicles will likely shift electricity demand, and the effect of upstream charging emissions will come from generation sources that are dispatched in response. This study proposes a multi-stage charging and discharging problem to translate low-cost energy transactions into vehicle dispatch decisions. A day-ahead charging optimization problem minimizes electricity purchases and marginal emissions damages, with energy transactions becoming targets in an optimization-based dispatch strategy for an on-demand shared autonomous electric vehicle (SAEV) fleet. The framework was tested for Austin, Texas, using an agent-based simulator. Fleets can schedule charging to lower daily power costs (averaging 15.5% or $0.79/day/SAEV) while reducing health damages from generation-related pollution (2.8% or $0.43/day/SAEV). Fleet managers can increase profits ($8 per SAEV per day) by adopting a multi-stage charging and discharging strategy that can serve more passengers per day than price-agnostic dispatch strategies.