RL-Core – Reinforcement Learning-based Competition in Reconfigurable Wireless Environments
This research project explores how reinforcement learning can enable future wireless communication systems to dynamically adapt to changing environments.
About RL-Core
Modern mobile networks increasingly rely on reconfigurable intelligent surfaces that can actively shape how radio signals propagate through the environment. These surfaces have the potential to improve signal quality, increase network capacity, and reduce energy consumption. At the same time, they introduce new challenges when multiple mobile network operators share the same wireless spectrum.
This project studies how reconfigurable intelligent surfaces affect competition and cooperation between network operators. A configuration that benefits one network may unintentionally degrade the performance of another, especially across different frequency bands. Without coordination, this can lead to conflicts and inefficient use of shared wireless resources.
To address this, the project develops coordination mechanisms based on reinforcement learning that allow operators to manage reconfigurable intelligent surfaces in a fair and efficient way. An independent broker supports negotiation between operators and learns, over time, how to balance network performance, user requirements, and energy efficiency.
The overall goal is to enable better wireless services with lower energy consumption, while ensuring fair access and reliable performance for users in shared network environments.
