§3, p. 362 — is a saddle point of iff is optimal for (P) and is a Kuhn–Tucker vector
ProvedPenaltyLag.Exact.saddle_iffLet be a nonempty convex subset of a real vector space and convex functions on , defining the convex program (P). Let and let be the penalty Lagrangian (2.3). For and , the pair is a saddle point of on , that is, and
if and only if is an optimal solution to (P) and is a Kuhn–Tucker vector for (P) relative to .
This is the penalty-Lagrangian counterpart of the classical saddle-point characterization of optimality, and the bridge used in the proof of Theorem 3.5.
Formalization Note The standing assumption of p. 358 (convex , convex ) is a hypothesis. The pair is arbitrary: if both sides fail. Saddle points are in the sense of Rockafellar's Convex Analysis, §36 (minimum over , maximum over , no sign restriction on ).
import Mathlib import Definitions.Def_PenaltyLag_Exact_Basic
namespace PenaltyLag.Exact
/-- §3, p. 362: for r > 0, (x̄, ȳ) is a saddle point of L_r iff x̄ is an optimal solution
to (P) and ȳ is a Kuhn–Tucker vector relative to L_r. -/
theorem saddle_iff {E : Type*} [AddCommGroup E] [Module ℝ E] {m : ℕ}
(X : Set E) (hX : Convex ℝ X) (hXne : X.Nonempty)
(f₀ : E → ℝ) (f : Fin m → E → ℝ) (hf₀ : ConvexOn ℝ X f₀) (hf : ∀ i, ConvexOn ℝ X (f i))
(r : ℝ) (hr : 0 < r) (xbar : E) (ybar : PenaltyLag.Asymptotic.Mult m) :
IsSaddle X f₀ f r xbar ybar ↔ (IsOptimal X f₀ f xbar ∧ IsKTVector X f₀ f r ybar) := by sorry
end PenaltyLag.Exact
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.