Proof of Theorem 2.1, p. 46 — contraction principle on V_ε with the bound (2.5): ‖x(p) − x‖ ≤ (1 − λδ)⁻¹‖Φ_p(x) − x‖
ProvedRobinsonSR.IFT.contraction_principle_ballLet be a real Banach space, , , and the closed ball of radius about . Let map into itself and satisfy, for some with ,
Then has a unique fixed point in , and for each
In the proof of Theorem 2.1 this is applied with and ; the fixed point is the solution , and (2.5) is the bound from which (2.4) follows.
Formalization Note is assumed complete. The page applies the contraction principle on in a space it calls only normed; on a closed ball of an incomplete space the principle can fail, so completeness is the hypothesis under which the principle, as invoked, holds. Theorem 2.1 itself is stated without completeness. The values of off play no role, and uniqueness is among fixed points in .
import Mathlib
namespace RobinsonSR.IFT
/-- Proof of Theorem 2.1, p. 46, the contraction principle as invoked there: a self-map `Φ` of
the closed ball `V_ε = closedBall x₀ ρ` (`ρ > 0`) of a Banach space that is a contraction with
modulus `κ ∈ [0, 1)` on it has a unique fixed point `x_p ∈ V_ε`, and
`‖x_p − x‖ ≤ (1 − κ)⁻¹ ‖Φ x − x‖` for every `x ∈ V_ε` (bound (2.5), with `κ = λδ`). -/
theorem contraction_principle_ball {X : Type*} [NormedAddCommGroup X] [NormedSpace ℝ X]
[CompleteSpace X] (x0 : X) (rho : ℝ) (hrho : 0 < rho) (Φ : X → X)
(hmaps : Set.MapsTo Φ (Metric.closedBall x0 rho) (Metric.closedBall x0 rho))
(kappa : ℝ) (hk0 : 0 ≤ kappa) (hk1 : kappa < 1)
(hcontr : ∀ x₁ ∈ Metric.closedBall x0 rho, ∀ x₂ ∈ Metric.closedBall x0 rho,
‖Φ x₁ - Φ x₂‖ ≤ kappa * ‖x₁ - x₂‖) :
∃ xp ∈ Metric.closedBall x0 rho, Φ xp = xp ∧
(∀ z ∈ Metric.closedBall x0 rho, Φ z = z → z = xp) ∧
∀ x ∈ Metric.closedBall x0 rho, ‖xp - x‖ ≤ (1 - kappa)⁻¹ * ‖Φ x - x‖ := by sorry
end RobinsonSR.IFT
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.