AI-Driven Deep Q-Learning Framework for Real-Time Electric Vehicle Charging Optimization and Smart Mobility Infrastructure Management

Abstract

The increasing adoption of electric vehicles (EVs) has created new challenges in managing charging infrastructure, optimizing energy consumption, and ensuring efficient mobility within smart city environments. Traditional charging methods often rely on static scheduling approaches that cannot effectively respond to dynamic traffic conditions, fluctuating electricity prices, varying charging demands, and grid load variations. To address these challenges, this paper proposes an AI-Driven Deep Q-Learning Framework for Real-Time Electric Vehicle Charging Optimization and Smart Mobility Infrastructure Management. The proposed framework utilizes Deep Q-Learning (DQL), a reinforcement learning algorithm, to make intelligent charging decisions by continuously learning from real-time environmental conditions. The model considers multiple state parameters, including battery state of charge, charging station availability, traffic congestion, electricity pricing, and grid load, to determine the most appropriate charging station and charging schedule. A reward function is designed to minimize charging waiting time, reduce energy costs, improve charging station utilization, and balance the electrical load across the network. The framework also supports adaptive decision-making, enabling continuous learning as traffic and energy conditions change. The proposed approach is expected to enhance charging efficiency, reduce operational costs, improve user convenience, and contribute to the sustainable management of smart mobility infrastructure. The study demonstrates the potential of integrating artificial intelligence and reinforcement learning into electric vehicle charging systems to support intelligent transportation networks and future smart city applications.