Deep Reinforcement Learning
Policy learning algorithms for complex, sequential decision environments, including model-free deep RL, domain adaptation, and sim-to-real transfer.
7 publications
Reader in Computer Science
Maria Chli
Aston University, Birmingham
Artificial intelligence for intelligent agents, multi-agent systems, and smart cities — from theoretical foundations to real-world deployment.
Research Themes
Intelligent Transportation
AI-driven approaches to urban traffic management — real-time signal optimisation, multi-intersection coordination, simulation environments (Traffic3D), and data synthesis with graph neural networks.
11 publications
Mechanism Design & Auctions
Game-theoretic frameworks for strategic procurement, resource allocation, and incentive compatibility, including optimal auction design for gradual service-provider engagement.
10 publications
Multi-Agent Systems
Coordination, competition, and learning in systems of intelligent agents — from formal models of rational behaviour to practical architectures for decentralised decision-making.
11 publications
Responsible AI
Co-design methodologies and stakeholder-inclusive frameworks for deploying AI systems that are fair, transparent, and beneficial — with a focus on sustainable urban mobility.
2 publications
Smart Cities & EV Infrastructure
Applying AI and game-theoretic optimisation to electric vehicle charging network design — strategic placement, pricing, and demand management at city scale.
3 publications
Research Projects
All projects →Mechanism Design for Strategic Agent Procurement
2019–ongoing
Optimal auction theory for the gradual, incentive-compatible procurement of strategic service-provider agents, with applications to outsourcing and platform economics.
AI-Optimised Electric Vehicle Charging Infrastructure
2022–ongoing
Game-theoretic models for optimising the placement and pricing of EV charging stations at city scale, balancing operator revenue, user welfare, and grid stability.
Responsible AI for Sustainable Urban Mobility
2023–ongoing
Participatory co-design of AI-powered travel recommendation systems that are transparent, fair, and aligned with the sustainability goals of diverse urban stakeholders.
Traffic3D: AI Traffic Simulation Platform
2018–2024
An open-source, photorealistic 3D traffic simulation environment for training and evaluating reinforcement learning agents under realistic urban traffic conditions.
Multi-Agent Deep RL for Traffic Signal Optimisation
2019–2023
Cooperative multi-agent reinforcement learning approaches to coordinating traffic signals across multiple road intersections, demonstrated on live camera feeds.