Fine capped upper first-crossing Brownian grids converge pathwise to the true stopped exponential
ProvedKyprianou.Ch01.brownianCappedFirstGridPointwiseConvergencePublicbrownian-motionfinite-gridfirst-passagekyprianou-ch01pathwise-convergence
For almost every continuous Brownian path, its compensated adjustment-root exponential evaluated at the earliest strict upper-crossing point of successively finer physical grids of [0,T] converges to the same exponential at the genuine continuous first upper-passage time capped by T. The grid retains the terminal sentinel and uses K=N+1; continuity, the infimum property and detection of strict-crossing neighbourhoods underpin the result.
Preamble
import Mathlib.MeasureTheory.Integral.Bochner.Basic import Mathlib.Analysis.SpecialFunctions.Exp import Definitions.Def_Kyprianou_Ch01_McKeanBrownianStopping import Definitions.Def_Kyprianou_Ch01_ExplicitLevyExamples open MeasureTheory ProbabilityTheory Kyprianou.Ch01 Filter open scoped NNReal ENNReal set_option autoImplicit false
Formal statement
theorem Kyprianou.Ch01.brownianCappedFirstGridPointwiseConvergencePublic
{Ω : Type*} [MeasurableSpace Ω]
(P : Measure Ω) [IsProbabilityMeasure P]
(X : ℝ≥0 → Ω → ℝ)
(sigma gamma q x : ℝ)
(hsigma : 0 < sigma) (hq : 0 < q) (hx : 0 < x)
(hX : IsLinearBrownian P X sigma gamma)
(hMeas : ∀ t : ℝ≥0, Measurable (X t))
(T : ℝ≥0) :
let theta : ℝ :=
(Real.sqrt (gamma ^ 2 + 2 * sigma ^ 2 * q) - gamma) / sigma ^ 2
let G : ℕ → Ω → ℝ := fun N ω =>
let K : ℕ := N + 1
let delta : ℝ≥0 := T / (K : ℝ≥0)
let tau : Ω → ℕ := fun ξ => by
classical
exact Nat.find (show ∃ k : ℕ,
k = K ∨ x < X ((k : ℝ≥0) * delta) ξ from ⟨K, Or.inl rfl⟩)
Real.exp (theta * X ((tau ω : ℝ≥0) * delta) ω -
q * (((tau ω : ℝ≥0) * delta) : ℝ))
let tauT : Ω → ℝ≥0 := fun ω =>
match firstUpperPassage X x ω with
| none => T
| some t => min t T
∀ᵐ ω ∂P, Tendsto (fun N : ℕ => G N ω) atTop
(nhds (Real.exp (theta * X (tauT ω) ω - q * (tauT ω : ℝ)))) := by sorrySource
Kyprianou (2006), Exercise 1.10(ii), back-of-book continuous first-passage and finite-grid approximation.