Proof of Proposition 1, pp. 6–7 — s²_n(x) → σ²_x̂(x) uniformly on X, w.p.1
ProvedSolutionQuality.SRP.sampleVar_tendstoUniformlyOnp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1sample-variancestochastic-programminguniform-law-of-large-numbers
Assume (A1)–(A3), let , and let be i.i.d. as . For let
be the sample variance of the differences , and . Then, with probability one,
This is the first step of the proof of Proposition 1 (iii): combined with part (ii) it controls the sample variance evaluated at the random minimizer .
Formalization Note. The minimizers do not appear. is assumed measurable for each . is Mathlib's variance, finite under (A2).
Preamble
import Mathlib import Definitions.Def_SolutionQuality_SRP_Setting
Formal statement
namespace SolutionQuality.SRP
open MeasureTheory ProbabilityTheory Filter
theorem sampleVar_tendstoUniformlyOn {d : ℕ} {Ξ : Type*} [MeasurableSpace Ξ]
(μ : Measure Ξ) [IsProbabilityMeasure μ]
(f : E d → Ξ → ℝ) (hfm : ∀ x, Measurable (f x))
(X : Set (E d)) (hA : Assumptions μ f X)
{Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(ξ : ℕ → Ω → Ξ) (hξ : IsIIDSample P μ ξ)
(xhat : E d) (hxhat : xhat ∈ X) :
∀ᵐ ω ∂P, TendstoUniformlyOn (fun n x => sampleVar f xhat (fun i => ξ i ω) n x)
(gapVar μ f xhat) atTop X := by sorry
end SolutionQuality.SRP
Source
Bayraksan & Morton, Assessing Solution Quality in Stochastic Programs, preprint (January 26, 2005), pp. 6–7, proof of Proposition 1 (iii), first claim
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.