Theorem 8.1.5 — continuous mapping for almost surely continuous functionals
OpenDurrettProbability.donsker_continuous_mappingLet be measurable, jointly independent, all with the law of , with
the square being integrable. Let be a Brownian motion on a second probability space, every one of whose paths is continuous, and write and for the associated random elements of .
Let be measurable, and suppose is continuous at almost every Brownian path: for almost every , is continuous at the point of . Then
This is the Mann–Wald continuous mapping theorem in the form that tolerates discontinuities, provided they form a set of limit-law measure zero — which is what makes it usable for the maximum, the occupation time, and the last zero.
Formalization Note Continuity is required only at the points where the limit law lives, and only almost everywhere with respect to it; may be wildly discontinuous elsewhere. Measurability of is assumed separately, since almost-everywhere continuity with respect to a measure does not by itself give Borel measurability of the map on the whole space.
Convergence in distribution is weak convergence of the push-forward laws, here of real random variables on possibly different probability spaces. The hypotheses on the walk are those of Durrett's Theorem 8.1.2, carried through unchanged: i.i.d. with mean and variance .
Assuming every Brownian path is continuous, rather than almost every one, is what makes a -valued random variable at all; it costs nothing, since a Brownian motion may be modified on a null set to achieve it.
import Mathlib import Definitions.Def_DurrettProbability_Brownian import Definitions.Def_DurrettProbability_Donsker open Filter MeasureTheory ProbabilityTheory open scoped NNReal Topology
namespace DurrettProbability
theorem donsker_continuous_mapping
{Ω : Type*} [MeasurableSpace Ω] {P : Measure Ω} [IsProbabilityMeasure P]
{X : ℕ → Ω → ℝ} (hmeas : ∀ k, Measurable (X k)) (hindep : iIndepFun X P)
(hident : ∀ k, IdentDistrib (X k) (X 0) P P) (hint : Integrable (fun ω => X 0 ω ^ 2) P)
(hmean : ∫ ω, X 0 ω ∂P = 0) (hvar : ∫ ω, X 0 ω ^ 2 ∂P = 1)
{Ω' : Type*} [MeasurableSpace Ω'] {P' : Measure Ω'} [IsProbabilityMeasure P']
{B : ℝ≥0 → Ω' → ℝ} (hB : IsBrownianReal B P') (hBc : ∀ ω, Continuous fun t => B t ω)
(ψ : PathSpace → ℝ) (hψm : Measurable ψ)
(hψ : ∀ᵐ ω ∂P', ContinuousAt ψ (brownianPath B ω)) :
TendstoInDistribution (fun n : ℕ => fun ω => ψ (walkPath X n ω)) atTop
(fun ω => ψ (brownianPath B ω)) (fun _ => P) P' := by sorry
end DurrettProbabilityConfirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.