Knowledge Distillation for Efficient Transformer-Based Reinforcement Learning in Hardware-Constrained Energy Management Systems
ACM SIGEnergy Energy Informatics Review · 01.09.2026 · 6 (3) · 138-152
DOI: 10.1145/3861130.3861141 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM SIGEnergy Energy Informatics Review, online erschienen am 11.11.2026. Verfasst von Pascal Henrich und 5 weiteren Autor:innen. Die Arbeit nennt 42 Literaturangaben.
Abstract (Originalsprache)
Transformer-based reinforcement learning has emerged as a strong candidate for sequential control in residential energy management. In particular, the Decision Transformer can learn effective battery dispatch policies from historical data, thereby increasing photovoltaic self-consumption and reducing electricity costs. However, transformer models are typically too computationally demanding for deployment on resource-constrained residential controllers, where memory and latency constraints are critical. The present paper investigates Knowledge Distillation to transfer the decision-making behaviour of high-capacity Decision Transformer policies to compact models more suitable for embedded deployment. Using the Ausgrid dataset, we train teacher models in an offline, sequence-based Decision Transformer framework on heterogeneous multi-building data. Given the trained models, we distil smaller student models by matching the teachers' actions, thereby preserving control quality while reducing model size. Across a broad set of teacher-student configurations, distillation largely preserves control performance and even yields small improvements of up to 1 %, while reducing the parameter count by up to 96 %, the inference memory by up to 90 %, and the inference time by up to 63 %. Beyond these compression effects, comparable cost improvements are also observed when distilling into a student of identical architectural capacity. Overall, our results show that Knowledge Distillation makes Decision Transformer control more applicable for residential energy management on resource-limited hardware.
Zitieren
APA
Henrich, P., Sievers, J., Beichter, M., Blank, T., Mikut, R., & Hagenmeyer, V. (2026). Knowledge Distillation for Efficient Transformer-Based Reinforcement Learning in Hardware-Constrained Energy Management Systems. ACM SIGEnergy Energy Informatics Review, 6(3), 138-152. https://doi.org/10.1145/3861130.3861141
Vancouver
Henrich P, Sievers J, Beichter M, Blank T, Mikut R, Hagenmeyer V. Knowledge Distillation for Efficient Transformer-Based Reinforcement Learning in Hardware-Constrained Energy Management Systems. ACM SIGEnergy Energy Informatics Review. 2026;6(3):138-152. doi:10.1145/3861130.3861141
BibTeX
@article{Henrich2026,
author = {Henrich, Pascal and Sievers, Jonas and Beichter, Maximilian and Blank, Thomas and Mikut, Ralf and Hagenmeyer, Veit},
title = {Knowledge Distillation for Efficient Transformer-Based Reinforcement Learning in Hardware-Constrained Energy Management Systems},
journal = {ACM SIGEnergy Energy Informatics Review},
year = {2026},
volume = {6},
number = {3},
pages = {138--152},
doi = {10.1145/3861130.3861141}
}
EUVIMED