Lemma 4.2, p. 125 — with ρχ ≥ 1 the minimizer is eventually bracketed by x₂ and x_e (without the printed bound on K)
DisprovedNelderMeadLD.Conv1D.lemma_4_2Let be strictly convex with bounded level sets , and let be its minimizer. Let the parameters satisfy (2.1) and in addition , and let be the one-dimensional Nelder–Mead run from a nondegenerate, ordered initial interval . Write for the expansion point and , .
Then there is a smallest integer such that
and for this the minimizer is bracketed by and : , the open interval with these endpoints.
The lemma is the first step of the convergence proof: within finitely many iterations the method finds an interval of uncertainty containing the minimizer.
Formalization Note The printed lemma also asserts . That bound is false: for , , , , the bound is , but , so . The bound is not stated here and no substitute bound is invented; Theorem 4.1 uses only the existence of . "Smallest" is stated as: the condition fails at every . The coefficients are constrained by the standing conditions (2.1).
import Mathlib import Definitions.Def_NelderMeadLD_Conv1D_Algorithm open Filter Topology
namespace NelderMeadLD.Conv1D
theorem lemma_4_2 (f : ℝ → ℝ) (hf : StrictConvexOn ℝ Set.univ f)
(hlev : ∀ μ : ℝ, Bornology.IsBounded {x | f x ≤ μ})
(xmin : ℝ) (hmin : ∀ y, f xmin ≤ f y)
(ρ χ γ σ : ℝ) (hpar : ParamsOK ρ χ γ σ)
(p0 : ℝ × ℝ) (h0 : IsStart f p0)
(hρχ : 1 ≤ ρ * χ) :
∃ K : ℕ, Bracketed ρ χ f (run f ρ χ γ σ p0 K) ∧
(∀ j < K, ¬ Bracketed ρ χ f (run f ρ χ γ σ p0 j)) ∧
min (run f ρ χ γ σ p0 K).2 (xe ρ χ (run f ρ χ γ σ p0 K)) < xmin ∧
xmin < max (run f ρ χ γ σ p0 K).2 (xe ρ χ (run f ρ χ γ σ p0 K)) := by sorry
end NelderMeadLD.Conv1D
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.