Context-Aware Optimal Resource Management in Electric Vehicle Smart2Charge

Sharif, Muddsair (2025) Context-Aware Optimal Resource Management in Electric Vehicle Smart2Charge. Doctoral thesis, Birmingham City University.

[thumbnail of Muddsair Sharif PhD Thesis_Final Version_Final Award Jan 2025.pdf]
Preview
Text
Muddsair Sharif PhD Thesis_Final Version_Final Award Jan 2025.pdf - Accepted Version

Download (8MB)

Abstract

This thesis presents a novel approach to optimizing electric vehicle (EV) charging systems through a context-aware framework powered by deep reinforcement learning (DRL). The research addresses critical challenges in the EV ecosystem, balancing the needs of multiple stakeholders including end-users, grid operators, fleet managers, and charging station operators. At its core, a Deep Q-Network (DQN) algorithm outperforms other state-of-the-art DRL methods in managing complex, multi-objective optimization scenarios.

This work advances the field by bridging theoretical DRL models with practical EV charging implementations, offering a framework that optimizes outcomes for multiple stakeholders while promoting sustainable transportation. Through the Smart2Charge application, the research demonstrates how context-aware solutions can enhance both user experience and environmental sustainability. The application integrates real-time data including grid conditions, user preferences, charging station availability, and environmental factors to optimize charging decisions. Comprehensive testing through simulations and real-world scenarios validates the system’s effectiveness and adaptability across diverse operating conditions.

The proposed system achieves a 15% increase in overall energy efficiency, 10% reduction in charging costs for EV owners, 20% decrease in grid strain, and 10% reduction in CO₂ emissions through optimal integration of renewable energy sources. These advancements significantly contribute to both user satisfaction and environmental sustainability. This research paves the way for more intelligent, user-centric, and environmentally conscious EV charging systems, marking a significant step towards sustainable urban mobility.

Item Type: Thesis (Doctoral)
Dates:
Date
Event
20 January 2025
Accepted
Uncontrolled Keywords: Deep Reinforcement Learning, Context-Aware, Smart Charging, Electric Vehicle, Multi-Modal, Multi-Objective, Resource Optimality, Resource optimisation, Multi-Agent DRL Model
Subjects: CAH10 - engineering and technology > CAH10-01 - engineering > CAH10-01-08 - electrical and electronic engineering
CAH11 - computing > CAH11-01 - computing > CAH11-01-05 - artificial intelligence
Divisions: Doctoral Research College > Doctoral Theses Collection
Faculty of Computing, Engineering and the Built Environment > College of Computing
Depositing User: Louise Muldowney
Date Deposited: 24 Jan 2025 16:30
Last Modified: 24 Jan 2025 16:30
URI: https://www.open-access.bcu.ac.uk/id/eprint/16096

Actions (login required)

View Item View Item

Research

In this section...