automated-modeling

StarOR proposes a framework that synergizes Monte Carlo Tree Search with test-time reinforcement learning for automated optimization modeling, achieving state-of-the-art performance across multiple benchmarks.