Deep Reinforcement Learning-Based Charging Pricing for Autonomous Mobility-on-Demand System

The autonomous mobility-on-demand (AMoD) system plays an important role in the urban transportation system. charging behavior of AMoD fleet becomes a critical link between and In this paper, we investigate strategic pricing scheme for station operators (CSOs) based on non-cooperative Stackelberg game framework. equilibrium investigates competition among multiple CSOs, explores nexus CSOs operator. proposed framework, responsive operator (order-serving, repositioning, charging) is formulated as multi-commodity network flow model to solve energy-aware traffic problem. Meanwhile, soft actor-critic multi-agent deep reinforcement learning algorithm developed framework while considering privacy-conservation constraints CSOs. A numerical case study with city-scale real-world data used validate effectiveness