University of Oldenburg FK II – Department for Computer Science Digitalized Energy Systems Group D-26111 Oldenburg
Secretary:
Meike Burke
Regina Knippenberg
Industriestraße 11, Room 0-014
+49 (0) 441 - 798 2878
+49 (0) 441 - 798 2756
Head
Prof. Dr.-Ing. Astrid Niesse
Industriestraße 11, Room 0-004
+49 (0) 441 - 798 2750
+49 (0) 441 - 798 2756
Details
DES-News
Publication on the environment design of optimal power flow environments for reinforcement learning applications
The publication “Learning the optimal power flow: Environment design matters” by Thomas Wolgast and Astrid Nieße was successfully published in mid-August in the top-class open access journal Energy and AI. The paper uses deep reinforcement learning to approximate the optimal power flow, one of the most important optimisation problems in the energy system. In particular, we work out the great importance of the environment design, i.e. the formulation as a reinforcement learning problem, as a decisive factor for the resulting performance.
The publication “Learning the optimal power flow: Environment design matters” by Thomas Wolgast and Astrid Nieße was successfully published in mid-August in the top-class open access journal Energy and AI. The paper uses deep reinforcement learning to approximate the optimal power flow, one of the most important optimisation problems in the energy system. In particular, we work out the great importance of the environment design, i.e. the formulation as a reinforcement learning problem, as a decisive factor for the resulting performance.
