§2, p. 4 — the multipliers kλ, kμ make Scaled-AKKT hold for every linear objective
ProvedStrictCQ.Scaled.scaled_akkt_linearconstraint-qualificationp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1sequential-optimality
Let , be a constraint system on and a feasible point. Suppose and satisfy for all , whenever , , and (2.4):
Then for every the Scaled-AKKT condition holds at for the linear objective .
In the paper this is witnessed by the constant sequence with multipliers and ; it is the construction that shows (2.3) is the weakest strict constraint qualification for Scaled-AKKT.
Preamble
import Mathlib import Definitions.Def_StrictCQ_Scaled_Setting import Definitions.Def_StrictCQ_Scaled_Conditions open Filter Topology open scoped InnerProductSpace
Formal statement
namespace StrictCQ.Scaled
theorem scaled_akkt_linear {n m p : ℕ} (C : Constraints n m p)
(xs : EuclideanSpace ℝ (Fin n)) (hxs : xs ∈ C.feasible)
(l : Fin m → ℝ) (u : Fin p → ℝ) (hu : ∀ j, 0 ≤ u j) (hcomp : ∀ j, C.g j xs ≠ 0 → u j = 0)
(hnorm : max ‖l‖ ‖u‖ = 1)
(h24 : ∑ i, l i • gradient (C.h i) xs + ∑ j, u j • gradient (C.g j) xs = 0)
(c : EuclideanSpace ℝ (Fin n)) :
C.ScaledAKKT (fun x => ⟪x, c⟫_ℝ) xs := by sorry
end StrictCQ.Scaled
Source
Andreani, Martínez, Ramos & Silva, Strict constraint qualifications and sequential optimality conditions for constrained optimization, Optimization Online 5197 (version of November 12, 2015), p. 4, §2 (last paragraph of §2)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.