Proof of Theorem 9, p. 20 — F semiconvex at x̄ w.r.t. Rⁿ and 0 ∈ ∂F(x̄) imply x̄ minimizes F over Rⁿ
ProvedMifflinSemismooth.Optimality.minimizer_of_zero_memglobal-minimizernonsmooth-optimizationp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1semiconvex
Let be semiconvex at with respect to , and suppose . Then is a global minimizer of :
In the proof of Theorem 9 this is applied to when (giving part (a)), to when (giving part (b)(i)), and to when and (giving part (b)(ii)).
Formalization Note The page states the step separately for and for ; it is stated once here for a generic . Semiconvexity is assumed only at (with respect to ), which is all the step uses.
Preamble
import Mathlib import Definitions.Def_MifflinSemismooth_Optimality_Setting
Formal statement
namespace MifflinSemismooth.Optimality
/-- Mifflin (1976), §5, proof of Theorem 9, p. 20: if `F` is semiconvex at `x̄` with respect to
`ℝⁿ` and `0 ∈ ∂F(x̄)`, then `x̄` minimizes `F` over `ℝⁿ`. -/
theorem minimizer_of_zero_mem {n : ℕ} (F : EuclideanSpace ℝ (Fin n) → ℝ)
(xbar : EuclideanSpace ℝ (Fin n)) (hF : MifflinSemismooth.Extremal.SemiconvexAt Set.univ F xbar)
(h0 : (0 : EuclideanSpace ℝ (Fin n)) ∈ MifflinSemismooth.Extremal.genGrad F xbar) :
∀ x, F xbar ≤ F x := by sorry
end MifflinSemismooth.Optimality
Source
Mifflin, Semismooth and semiconvex functions in constrained optimization, IIASA Research Report RR-76-21 (December 1976), p. 20, §5, proof of Theorem 9, first two sentences
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.