Reproduce “RL-PaO: Prediction as Action in Decision Making under Uncertainty” 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
Jiahui Feng, Dafang Zhao, Zheng Chen, Zhengmao Li, Lingwei Zhu
RL-PaO trains a prediction policy using realized scheduling cost and prediction error as feedback. It predicts the per-period photovoltaic generation upper bound, which a mixed-integer linear programming solver uses to create a schedule; execution under realized conditions supplies the cost. On the 2019 test year, it incurred annual cost 1117.10 × 10⁴ versus 1207.90 × 10⁴ for supervised predict-and-optimize, in one offline scheduling application.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.