This paper introduces a data-driven bilevel framework for residential demand response that combines flexibility estimation from smart meter data with aggregator-household optimization under behavioral uncertainty. The approach models both automated and manual household participation, enabling 20.7 MWh of load shifting and EUR 1,029 in savings across 12 German households. Results highlight participation frequency as a key driver of economic performance and demonstrate the framework’s potential for scalable coordination of distributed residential flexibility.

Authors: Reda El Makroum, Sebastuab Zwickl-Bernhard, Lukas Kranzl, Hans Auer

The paper was published in April 2026.