Joint optimization of fleet sizing, charging station planning, and operation for autonomous electric vehicle fleets in urban transportation networks

The rapid advancement of autonomous electric vehicles (AEVs) is reshaping urban transportation, presenting new opportunities for ride-hailing fleets. Under a vertically integrated structure and a unified economic objective, this study develops a mixed-integer linear programming model that jointly optimizes vehicle loading/rebalancing, order acceptance/abandonment, and charging/discharging operations, while accounting for AEV and charging station investments, charging and maintenance costs, and discharging and passenger revenues to maximize the operator's net present value. One year of New York City Yellow Taxi trip data is processed by sampling from empirical distributions and introducing distributional noise using maximum likelihood estimation and the Akaike Information Criterion to capture travel demand characteristics and uncertainties. The model is tested across six operational modes within a 24-node transportation network that aggregates New York City's taxi pick-up and drop-off zones and integrates real-world travel distances, speeds, and electricity prices. Results show that rebalancing reduces investment costs by approximately 50 %, while a flexible order acceptance strategy strategically abandons extreme congestion orders during peak hours, resulting in a 3.5 % cost reduction. Discharging operations improve charging pile utilization by 7 %. Sensitivity analysis reveals that higher driving speed and vehicle-to-grid incentives enhance profitability, a charging power of charging piles with 80 kW achieves a favorable cost–benefit trade-off, while the marginal benefits of increasing AEV battery capacity and charging/discharging power gradually decline as operational benefits approach saturation. These findings offer a practical framework for operators and planners to deploy cost-effective AEV fleets in urban transportation networks.