The aim of this master thesis is to conduct a comparative analysis of algorithms for identifying control interfaces in power grids—specifically, boundaries where different grid segments connect with minimal or no actuator conflicts. Detecting these clear control interfaces can simplify the control task and improve the performance of reinforcement learning (RL) agents. This thesis will explore and compare methods from graph theory, optimization, and heuristic approaches to evaluate how effectively they pinpoint these interfaces under diverse grid conditions and switching states.

The identified control interfaces will be incorporated into an RL control framework, allowing a comparison between operating with aggregated control spaces (using the detected interfaces) and unsegmented grids. Key performance metrics such as convergence speed, stability, and overall control efficiency will be analyzed. The insights from this study are expected to contribute to more efficient and robust RL-based strategies for power grid management.

This thesis will be done in cooperation with the OFFIS Insitute. I will forward your applications to the colleague working there who will supervise your thesis. You can also reach him directly by writing an e-mail to sharaf.aldin.alsharif@offis.de

• Compare different methods to identify control interfaces—boundaries where grid segments connect with minimal actuator conflicts.
• Integrate the identified control interfaces into a reinforcement learning (RL) framework.
• Evaluate performance metrics in various grid conditions and switching states.
   

Proficiency in programming

Helpful knowledge, but not essential:
•    Python
•    Familiarity with power systems
•    Basic knowledge of reinforcement learning
•    Experience with simulation and modeling of complex systems