You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Code and data for "Do Language Models Know Which Constraints Belong Together?": zero-marginal groups in Dantzig-Wolfe reformulation, the JointDW search, and the experiments on MIPLIB-NL, rich models and benchmark instances.
Log-driven LLM tuning of SCIP for mixed-integer optimization, with solver-log analysis, scalar-feedback baselines, and a reproducible 215-instance benchmark.