Skip to main content
Search

Amirhosein Moshari

1 related articles

Subscribe to updates
  1. Lund thesis matches reinforcement learning, a KNN surrogate, and an LLM semantic workflow to three data conditions across the building life cycle

    This thesis develops three complementary AI approaches matched to the data available at different building life-cycle stages: a model-free reinforcement learning supervisory heating controller deployed in thirteen occupied buildings in Austria and evaluated over 138 days, a KNN surrogate trained on four years of Swedish residential measurements for safe offline evaluation of control strategies, and a multi-agent semantic workflow combining large language models and Sentence-BERT to interpret IFC/BIM material descriptions and link them to validated thermal, mass, and energy-related properties.

You’re all caught up