Summary
Electric vehicles don’t emit pollution while driving, but how clean their charging is depends on the energy mix powering the grid. The authors study how to plan routes and charging times for electric vehicles to reduce the carbon emissions from electricity production by charging when cleaner energy is available. They develop a special way to train prediction models that focus on helping these routing decisions rather than just making the most accurate forecasts. Their experiments on German data show that this approach reduces emissions by nearly 29% compared to ignoring carbon emissions. They also show their method leads to more cost-efficient routes than standard prediction models, even though the predicted carbon levels are less accurate.
What this means in practice
- •For fleet managers: Optimize electric vehicle routes and charging schedules to reduce carbon emissions based on grid clean energy forecasts.
- •For energy grid operators: Incorporate vehicle charging demand shifts into grid load planning by leveraging carbon-aware routing predictions.
Authors
Jaume Reixach, Wanru Chen, Jiaming Wu, Balázs Kulcsár
Abstract
Electric vehicles produce no tailpipe emissions, but charging them produces emissions that depend on the time-varying generation mix of the electricity consumed. We study an Electric Vehicle Routing Problem with a carbon-aware objective that accounts for charging-related emissions alongside conventional routing costs. The problem determines routing, charging, and waiting decisions, allowing charging demand to be shifted toward periods with lower grid carbon intensity. Because these decisions depend on forecasts of future carbon intensity, we propose a decision-focused learning approach so that the training of the model employed for predicting the carbon intensity of the grid over time is solver-aware. In particular, we train a linear predictor using approximate SPO+ directions and employ a time-limited Large Neighborhood Search followed by exact postprocessing as the underlying solver. Experiments on realistic Germany-based instances show that explicitly considering charging-related emissions reduces realized emissions by 28.76% on average relative to carbon-blind optimization, although at the expense of higher non-carbon operating cost. Moreover, decision-focused training reduces realized total cost by 2.30% on average relative to conventional prediction-focused learning and performs better on 75.5% of the test instances, despite substantially lower pointwise predictive accuracy. Finally, we show that its average advantage remains positive across varied solver configurations and carbon-cost weights.