We, the junior research group, RL4CES - Reinforcement Learning for Cognitive Energy Systems,aim to realize the potential of Deep Reinforcement Learning (Link DRL) in the context of the energy system.

With our Research we seek to make Deep Reinforcement Learning safer, more effective and more cost-effective for the energy industry and thus make a significant contribution to a stable and resilient energy system.

The focus of our application is on the topics of automated grid control and automated energy trading.

As the proportion of renewable energies in the European energy grid increases, its complexity is also growing significantly.

Unpredictable weather fluctuations are transmitted directly to the power grid by wind or solar energy. As a result, human expertise alone is no longer sufficient to cope with the increasingly complex tasks of the energy industry.

Intelligent, automated solutions need to be developed to prevent the real risk of localised or even widespread power outages.

Deep reinforcement learning has achieved some promising results in recent years in areas such as XXX. It is ideally suited for use in complex, dynamic systems such as the energy system, but - like any form of machine learning - it brings its own challenges for each field of application.

We meet these challenges through scientifically sound, practical Research.

We are a joint junior research group of the Fraunhofer Institute for Energy Economics (IEE) and the Department of Intelligent Embedded Systems (IES) at Kassel University.

The IES department is actively engaged in both basic and applied research, particularly delving into intelligent technical systems. Presently, their efforts are concentrated on analyzing time series data and exploring active learning methodologies. In addition to application areas such as transport and materials science, use cases in the field of renewable energies and future energy systems have been realised here.
In addition, the GAIN junior research group is investigating issues relating to graph neural networks in a further collaboration with the Fraunhofer IEE.
In the case of energy systems, the IES department is currently involved in the TRANSFER, Digital Twin Solar, SyLasKI and SALM projects, which focus, for instance, on transfer learning for wind and PV forecasts (Transfer), anomaly detection, active learning and probabilistic models in digital twins of photovoltaic systems (Digital Twin Solar).

The IES department specializes in targeted theoretical groundwork for challenges within the energy industry, leveraging its expertise to develop robust theoretical frameworks. Meanwhile, Fraunhofer IEE's cutting-edge IT infrastructure, simulation environments, and deep reinforcement learning (DRL) proficiency provide the essential foundation for highly specialized advanced research endeavors.
The close collaboration between IEE and IES is designed to ensure seamless continuity from fundamental research to practical application, specifically incorporating reinforcement learning techniques into the energy industry.

This symbiotic relationship will not only facilitate concurrent progress in research and application within the project but also foster the growth of a new cohort of experts. Through their doctoral studies, these individuals will receive invaluable guidance and mentorship from both Fraunhofer IEE and the IES department at the University of Kassel, enriching both their personal and scientific development.