Theorem 8, p. 19 — semiconvex F on a convex X: F(x+d) ≦ F(x) implies F'(x;d) ≦ 0
ProvedMifflinSemismooth.Optimality.theorem8directional-derivativenonsmooth-analysisp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1semiconvex
Let be convex and let be semiconvex on (at each point of , with respect to ). Let and with . Then the directional derivative exists, and
This is a nonsmooth analogue of the fact that a pseudoconvex function cannot increase infinitesimally in a direction along which it does not increase overall. In the proof of Theorem 9 it is applied to the constraint function at a point with .
Formalization Note is the relation HasDirDeriv; the conclusion states that the limit exists and that every value it has is . Existence is part of Definition 2(b) at , so the conjunct adds nothing beyond the page.
Preamble
import Mathlib import Definitions.Def_MifflinSemismooth_Optimality_Setting
Formal statement
namespace MifflinSemismooth.Optimality
/-- Mifflin (1976), §5, Theorem 8, p. 19: if `F` is semiconvex on a convex set `X`, `x ∈ X` and
`x + d ∈ X`, then `F(x + d) ≤ F(x)` implies `F'(x; d) ≤ 0`. -/
theorem theorem8 {n : ℕ} (F : EuclideanSpace ℝ (Fin n) → ℝ) (X : Set (EuclideanSpace ℝ (Fin n)))
(hX : Convex ℝ X) (hF : MifflinSemismooth.Extremal.SemiconvexOn X F) (x d : EuclideanSpace ℝ (Fin n)) (hx : x ∈ X)
(hxd : x + d ∈ X) (hle : F (x + d) ≤ F x) :
(∃ L, MifflinSemismooth.Extremal.HasDirDeriv F x d L) ∧ ∀ L, MifflinSemismooth.Extremal.HasDirDeriv F x d L → L ≤ 0 := by sorry
end MifflinSemismooth.Optimality
Source
Mifflin, Semismooth and semiconvex functions in constrained optimization, IIASA Research Report RR-76-21 (December 1976), p. 19, Theorem 8
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.