Theorem 2.10 — iterates of a continuous -Fejér map converge to a point of
ProvedShorNonsmooth.Fejer.fejer_map_iterates_convergefejer-mapfixed-point-iterationp2o-batch-b23ap2o-gran-per-chapterp2o-plan-bookp2o-v1
Let be nonempty and let be an -Fejér map: and for all and . If is continuous, then for every starting point the sequence
converges to an element of : there is with .
This is the basic convergence principle of Fejér-type approximations (Eremin). Polyak's method is built on the same inequality, but its map need not be continuous, so the principle does not apply to it directly.
Formalization Note is any nonempty subset of (no closedness is assumed, as in the book). The sequence is any satisfying the recursion; its first term is arbitrary.
Preamble
import Mathlib import Definitions.Def_ShorNonsmooth_Fejer_PolyakMethod open Filter Topology
Formal statement
namespace ShorNonsmooth.Fejer
/-- Shor (1985), p. 37, Theorem 2.10 (after Eremin [23]). If an `M`-Fejér map `φ` is
continuous, then for any starting point `x₀` the sequence generated by `x_{k+1} = φ(x_k)`,
`k = 0, 1, …`, converges to an element of `M`. -/
theorem fejer_map_iterates_converge {n : ℕ} (M : Set (EuclideanSpace ℝ (Fin n)))
(φ : EuclideanSpace ℝ (Fin n) → EuclideanSpace ℝ (Fin n))
(hφ : IsMFejer M φ) (hcont : Continuous φ)
(x : ℕ → EuclideanSpace ℝ (Fin n)) (hx : ∀ k, x (k + 1) = φ (x k)) :
∃ xbar ∈ M, Tendsto x atTop (𝓝 xbar) := by sorry
end ShorNonsmooth.Fejer
Source
Shor, Minimization Methods for Non-Differentiable Functions, Springer 1985, p. 37, Theorem 2.10 (with Definition (2.31), p. 36)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.