Pre-Day Scheduling of Charging Processes in Mobility-on-Demand Systems
Considering Electricity Price and Vehicle Utilization Forecasts
Abstract
Electrifying mobility-on-demand (MoD) fleets is
an important step towards a more sustainable transportation
system. With increasing fleet size, MoD operators will be
able to participate in the energy exchange market and will
have access to time-varying electricity prices. They can benefit from
intelligent scheduling of charging processes considering forecasts of
electricity prices and vehicle utilization. Considering a long time
horizon of, e.g., a day improves scheduling decisions, but electricity
prices change in a short interval of 15 minutes; hence, an
optimization-based approach needs to overcome challenges regarding
computational time. For this reason, we develop a macroscopic model to
study the tradeoffs between electricity, battery wear and
level-of-service costs. In scenarios with varying fleet size and
different numbers of
charging units, we compare the performance of several reactive and
scheduling policies in a simulation framework based on a macroscopic
model. Overall, the results of the study show that an MoD provider with
2000 vehicles could save several thousands of euros in daily operational
costs by changing from a state of charge reactive charging strategy to
one adapting to the price fluctuations of the electricity exchange
market.