Reproduce “Learning to Optimise Wind Farms with Graph Transformers” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Siyi Li, Arnaud Robert, A. Aldo Faisal, Matthew D. Piggott
The transformer predicts turbine power from farm layouts, yaw angles, and wind conditions, letting a genetic algorithm evaluate candidates in batches. It uses connected attention to learn turbine interactions and is trained on PyWake simulations. On a 55-turbine multi-string layout, its selected setting reached 99.3% relative accuracy and took about one minute versus 30.4 minutes for PyWake, not field validation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.