Proof of Theorem 2.4, p. 3 — h(Xₙ) = (Ȳₙ(1) − μ)/g(Ȳₙ) by (2.3i) and (2.3ii)
ProvedGlynnSTS.Limit.hmap_Xn_eqhomogeneityp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1standardized-time-series
Let satisfy (2.3i), for , and (2.3ii), for , where . Let for , otherwise, and for an arbitrary -valued and . Then for every and every ,
This converts the continuous-mapping step into the statement of Theorem 2.4: is the standardized time average.
Formalization Note is required because (2.3i) is applied with , which must be positive. Only (2.3i) and (2.3ii) are assumed; the identity is algebraic and holds for every , so Assumption (2.1) is not needed. Where both sides are .
Preamble
import Mathlib import Definitions.Def_GlynnSTS_Limit_Setting open MeasureTheory ProbabilityTheory Filter Topology open scoped NNReal ENNReal
Formal statement
namespace GlynnSTS.Limit
theorem hmap_Xn_eq {Ω : Type*} (Ybar : ℕ → Ω → C(unitInterval, ℝ)) (μ : ℝ)
(g : C(unitInterval, ℝ) → ℝ)
(hhom : ∀ a : ℝ, 0 < a → ∀ x, g (a • x) = a * g x)
(hshift : ∀ β : ℝ, ∀ x, g (x - β • kfun) = g x) :
∀ n : ℕ, 1 ≤ n → ∀ ω, hmap g (Xn Ybar μ n ω) = (Ybar n ω 1 - μ) / g (Ybar n ω) := by sorry
end GlynnSTS.Limit
Source
Glynn & Iglehart, Simulation output analysis using standardized time series, Math. Oper. Res. 15 (1990), p. 3, proof of Theorem 2.4
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.