Lemma 3.5 (case n = 1), p. 122 — for strictly convex f no shrink step is taken by the 1-D Nelder–Mead method
ProvedNelderMeadLD.Conv1D.lemma_3_5_one_dimnelder-meadp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1strict-convexity
Let be strictly convex, let the parameters satisfy (2.1), and run Algorithm NM in dimension from a nondegenerate initial interval , , ordered so that . Then for every iteration ,
This is the one-dimensional case of the paper's Lemma 3.5. It is why the shrink coefficient plays no role in the one-dimensional analysis.
Formalization Note The general- Lemma 3.5 belongs to a separate mission of this series; this item is its case on the one-dimensional algorithm of this mission. Bounded level sets are not needed and not assumed.
Preamble
import Mathlib import Definitions.Def_NelderMeadLD_Conv1D_Algorithm open Filter Topology
Formal statement
namespace NelderMeadLD.Conv1D
theorem lemma_3_5_one_dim (f : ℝ → ℝ) (hf : StrictConvexOn ℝ Set.univ f)
(ρ χ γ σ : ℝ) (hpar : ParamsOK ρ χ γ σ)
(p0 : ℝ × ℝ) (h0 : IsStart f p0) :
∀ k : ℕ, move f ρ χ γ (run f ρ χ γ σ p0 k) ≠ Move.shrink := by sorry
end NelderMeadLD.Conv1D
Source
Lagarias, Reeds, Wright & Wright, Convergence properties of the Nelder–Mead simplex method in low dimensions, SIAM J. Optim. 9 (1998), p. 122, Lemma 3.5 (case n = 1)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.