Characteristic function of Brownian first-passage time from the principal complex square-root Laplace boundary
ProvedKyprianou.Ch01.brownianFirstPassageCharacteristicComplexSqrtPublicbrowniancharacteristic-functioncomplex-laplacekyprianou-ch01stable-half
Let T be the first positive-barrier upward crossing time of centred unit-variance Brownian motion, set to zero if no crossing. For every real frequency xi, show that the characteristic function of the pushforward distribution of T equals exp(-s*sqrt(-2 i xi)), with the principal complex square root. This is the analytic-continuation boundary theorem for the already proved real Laplace transform E[exp(-qT)]=exp(-s sqrt(2q)); justify continuity and dominated passage to imaginary q and measurability of the finite stopping time. This theorem does not assume the canonical stable skew exponent and does not use the target stable-parameters theorem.
Preamble
import Definitions.Def_LeanLevy_Processes_LevyProcess import Definitions.Def_Kyprianou_Ch01_McKeanRightContinuousFiltration import Mathlib import Definitions.Def_Kyprianou_Ch01_RiskQueueCore import Definitions.Def_Kyprianou_Ch01_ExplicitLevyExamples import Definitions.Def_Kyprianou_Ch01_McKeanBrownianStopping open MeasureTheory ProbabilityTheory Kyprianou.Ch01 Set Filter Complex open scoped NNReal ENNReal set_option autoImplicit false universe u
Formal statement
theorem Kyprianou.Ch01.brownianFirstPassageCharacteristicComplexSqrtPublic
{Ω : Type u} [MeasurableSpace Ω]
(P : Measure Ω) [IsProbabilityMeasure P]
(B : ℝ≥0 → Ω → ℝ)
(hB : Kyprianou.Ch01.IsLinearBrownian P B 1 0)
(hMeas : ∀ t : ℝ≥0, Measurable (B t))
(hContinuous : ∀ ω, Continuous (fun t : ℝ≥0 => B t ω))
(s : ℝ) (hs : 0 < s) :
let T : Ω → ℝ :=
fun ω => match Kyprianou.Ch01.firstUpperPassage B s ω with
| none => 0
| some t => (t : ℝ);
∀ ξ : ℝ,
charFun (P.map T) ξ =
Complex.exp
(-(s : ℂ) * Complex.sqrt ((-2 : ℂ) * Complex.I * (ξ : ℂ))) := by sorrySource
Kyprianou (2006), Exercise 1.10(ii), half-stable first-passage distribution via analytic continuation of Brownian Laplace transform.