# Learning to Run a Power Network under Varying Grid Topology

Source: IEEE ENERGYCON metadata via DOI/search result; ResearchGate/ADS metadata checked; arXiv search returned no arXiv version.
Retrieved: 2026-06-14

## Bibliographic Metadata

- Authors: Shams Taha; Jan Poland; Katarina Knezovic; Dmitry Shchetinin
- Venue/status: 2022 IEEE 7th International Energy Conference (ENERGYCON), pp. 1-6. DOI 10.1109/ENERGYCON53164.2022.9830198.

## Source Links

- https://doi.org/10.1109/ENERGYCON53164.2022.9830198
- https://www.researchgate.net/publication/362181678_Learning_to_Run_a_Power_Network_under_Varying_Grid_Topology
- https://ui.adsabs.harvard.edu/abs/2022ener.conf...66T/abstract

## Abstract / Retrieval Summary

Metadata-only entry for the Grid2Op/L2RPN paper cited by later surveys as a model-based graph RL approach: a learned graph convolutional predictor estimates line-loading outcomes under candidate actions/topologies and an MCTS planner uses that learned physics model to select action sequences. This is relevant as an early action-conditioned learned surrogate/world-model precedent rather than a current SOTA benchmark.

## Retrieval Note

No arXiv version or open standalone full-text PDF was found during targeted search. IEEE full-text access was not available from this environment. Stored as metadata-only.