Smart Grid-Driven Intelligent EV Charging Scheduling Strategy Based on Deep Q-Networks

With the increasing popularity of electric vehicles (EVs) and the promotion of environmental policies, the rational scheduling of EV charging behavior has become increasingly important to ensure both user needs and grid stability. This paper proposes an intelligent charging scheduling strategy based on a Deep Q-Network (DQN) to address the energy management and scheduling issues of electric vehicles. This strategy leverages deep reinforcement learning techniques to dynamically adjust charging and discharging behavior based on real-time electricity price data, aiming to maximize economic benefits and optimize grid load. Compared to traditional Q-Learning algorithms, DQN demonstrates significant advantages in handling high-dimensional state and action spaces. Our experimental results indicate that the DQN-based scheduling strategy can effectively reduce charging costs while responding to grid demands and optimizing energy usage efficiency under real-world electricity price fluctuations.