Kronqvist & Misener, OptEng 22:1315-1345 (2021), https://doi.org/10.1007/s11081-020-09551-6 (Rosenbrock Prize) — strengthen OA cuts using the disjunctive structure, solvable by enumeration, usable for pruning. No open-source implementation exists anywhere (verified Aug 2026) — first-mover opportunity, and slide 17 conjectured GDP structure does better than the MINLP version.
In GDPopt today: OA/affine cuts land on the owning Disjunct and get big-M'd via util._add_bigm_constraint_to_transformed_model using global variable bounds — no tightened M from disjunctive bounds (needs Pyomo#4014), no lifting, no CGLP-style strengthening, no cut pool management.
Scope: strengthen per-disjunct cuts à la Kronqvist-Misener enumeration; tightened-M placement once Pyomo#4014 lands; compare with Trespalacios & Grossmann multiple-M cuts.
Part of roadmap epic #2 (item B6).
Kronqvist & Misener, OptEng 22:1315-1345 (2021), https://doi.org/10.1007/s11081-020-09551-6 (Rosenbrock Prize) — strengthen OA cuts using the disjunctive structure, solvable by enumeration, usable for pruning. No open-source implementation exists anywhere (verified Aug 2026) — first-mover opportunity, and slide 17 conjectured GDP structure does better than the MINLP version.
In GDPopt today: OA/affine cuts land on the owning Disjunct and get big-M'd via
util._add_bigm_constraint_to_transformed_modelusing global variable bounds — no tightened M from disjunctive bounds (needs Pyomo#4014), no lifting, no CGLP-style strengthening, no cut pool management.Scope: strengthen per-disjunct cuts à la Kronqvist-Misener enumeration; tightened-M placement once Pyomo#4014 lands; compare with Trespalacios & Grossmann multiple-M cuts.
Part of roadmap epic #2 (item B6).