Temporal RL for Energy System Control

The temporal RL agent embeds temporal awareness to RL by integrating forecasts (e.g., load, PV, and wind generation) into decision-making through an attention-based temporal embedding module.

Citation

[1] A. H. Ardakani, J. Hurink, I. Gibson and E. Shirazi, Attention-Based Temporal Reinforcement Learning for Energy System Control, 2025 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT Europe), Valletta, Malta, 2025, pp. 1-5, doi: 10.1109/ISGTEurope64741.2025.11305619.

[2] A. H. Ardakani, J. Hurink, I. Gibson and E. Shirazi, Embedding Temporal Awareness in Reinforcement Learning Models for Energy System Control, Proceedings of the 16th ACM International Conference on Future and Sustainable Energy Systems (E-Energy '25). Association for Computing Machinery, New York, NY, USA, 2025, 1002–1004. https://doi.org/10.1145/3679240.3734686

Results

Inference

Run the following notebook:

$ 6. attention_TRL.ipynb

Data Sources

Data	Source
PV & Wind Generation	ENTSO-E Transparency Platform
Load	ENTSO-E Transparency Platform
Electricity Price	ENTSO-E Transparency Platform
CO2 Emission	Nationaal Energie Dashboard