Proposition 3.5 — a priori estimate
DisprovedReflectedBSDE.Existence.proposition_3_5For every horizon and Lipschitz constant there is a constant , uniform in the Brownian dimension , with the following property. Let satisfy (i)–(iv) with Lipschitz constant , and . Let be a square-integrable solution of the reflected BSDE, i.e. one satisfying (v), (v′) and (vi)–(viii). Then
The estimate bounds the size of the solution by the size of the data. It is the basic tool for the contraction argument behind existence and for the stability estimate (Proposition 3.6).
Formalization Note The paper says "there exists a constant " after fixing a solution. Read that way the statement is trivial, since the left side is finite. Its proof gives a constant that depends only on the Lipschitz constant and , so the formal statement puts before the dimension, probability space, data and solution. Expectations are lower Lebesgue integrals in . is the Euclidean norm, and the Lipschitz constant is called L in Lean.
import Mathlib import Definitions.Def_Peng1990_SMP_Stochastic import Definitions.Def_ReflectedBSDE_Existence_Skorohod import Definitions.Def_ReflectedBSDE_Existence_Setting import Definitions.Def_ReflectedBSDE_Existence_Solution open MeasureTheory ProbabilityTheory Set open scoped NNReal ENNReal open Peng1990.SMP
namespace ReflectedBSDE.Existence
universe u
/-- Proposition 3.5 (p. 709): a priori estimate. For every dimension `d`, horizon `T` and
Lipschitz constant `L` there is a constant `C`, uniform in the Brownian dimension, such that every square-integrable solution
`(Y, Z, K)` (conditions (v)–(viii) and (v′)) of the RBSDE with data `(ξ, f, S)` satisfying
(i)–(iv) with Lipschitz constant `L` obeys
`E(sup_{0≤t≤T} Y_t² + ∫₀ᵀ |Z_t|² dt + K_T²) ≤ C E(ξ² + ∫₀ᵀ f²(t, 0, 0) dt + sup_{0≤t≤T} (S_t⁺)²)`. -/
theorem proposition_3_5 :
∀ (T : ℝ≥0) (L : ℝ), ∃ C : ℝ≥0,
∀ (d : ℕ) (Ω : Type u) [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(B : ℝ≥0 → Ω → Fin d → ℝ) (hB : IsStdBrownian P B)
(ξ : Ω → ℝ) (f : ℝ≥0 → Ω → ℝ → (Fin d → ℝ) → ℝ) (S : ℝ≥0 → Ω → ℝ)
(Y : ℝ≥0 → Ω → ℝ) (Z : ℝ≥0 → Ω → Fin d → ℝ) (K : ℝ≥0 → Ω → ℝ),
IsStandardData (augmentedFiltration P hB) P T L ξ f S →
SatisfiesV (augmentedFiltration P hB) P T Z →
SolvesRBSDE (augmentedFiltration P hB) P T B ξ f S Y Z K →
SatisfiesVPrime (augmentedFiltration P hB) P T Y K →
∫⁻ ω, ((⨆ t ∈ Iic T, ‖Y t ω‖ₑ ^ 2)
+ (∫⁻ s in Icc (0 : ℝ) T, ENNReal.ofReal (euclSq (Z s.toNNReal ω)))
+ ‖K T ω‖ₑ ^ 2) ∂P
≤ (C : ℝ≥0∞) * ∫⁻ ω, (‖ξ ω‖ₑ ^ 2
+ (∫⁻ s in Icc (0 : ℝ) T, ‖f s.toNNReal ω 0 0‖ₑ ^ 2)
+ ⨆ t ∈ Iic T, ‖max (S t ω) 0‖ₑ ^ 2) ∂P := by sorry
end ReflectedBSDE.Existence
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.