Advancements in Electric Vehicle Charging Optimization: A Survey of Reinforcement Learning Approaches

TL;DR

This survey reviews reinforcement learning methods for optimizing electric vehicle charging in smart grids, highlighting recent advances, challenges, and future research directions to improve efficiency and stability.

Contribution

It provides a comprehensive classification and analysis of reinforcement learning approaches for EV charging, emphasizing recent developments and future research opportunities.

Findings

Reinforcement learning effectively manages EV charging in smart grids.

Deep learning enhances model-free, real-time optimization.

Future research should focus on decentralized methods and scalability.

Abstract

In response to global warming and energy shortages, there has been a significant shift towards integrating renewable energy sources, energy storage systems, and electric vehicles. Deploying electric vehicles within smart grids offers a promising solution to reduce carbon emissions. However, managing the charging and discharging processes of them as distributed power supplies present significant challenges. Additionally, the intermittent nature of renewable energy, uncertainties in electric vehicle-related parameters, fluctuating energy prices, and varying loads make maintaining stable power system operations more complex. Effective management systems for electric vehicle battery charging are crucial to coordinating these processes and ensuring a secure, efficient, and reliable power system. Reinforcement learning, enhanced by deep learning, has gained substantial interest for its model-free approach and real-time optimization, effectively managing electric vehicle charging by maximizing cumulative rewards. This review synthesizes existing literature on reinforcement learning-based frameworks, objectives, and architectures for electric vehicle charging coordination strategies in power systems, classifying methods into centralized and decentralized categories. Additionally, the article offers suggestions for future research directions to further enhance reinforcement learning-based electric vehicle charging optimization.