Example 1 — the constant mean satisfies β(1 − T) = E(J₀ + Y)
ProvedAntonelliBFSDE.SingularExample.example_1_betabackward-stochastic-differential-equationmartingalep2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
Let solve (3.8)–(3.9), and let be the mean of the martingale obtained by adding the two equations:
Then
This is the identity that rules out a solution at and, together with positivity, at larger horizons. The here denotes a martingale mean, not the bound on total variation used elsewhere in the paper.
Formalization Note is the displayed integral computed from the data rather than an arbitrary parameter. Strict positivity of is explicit. The conditional-expectation arguments are required integrable by the solution concept, avoiding a default-value interpretation.
Preamble
import Mathlib import Definitions.Def_AntonelliBFSDE_SingularExample_Setting open MeasureTheory open scoped NNReal ENNReal
Formal statement
namespace AntonelliBFSDE.SingularExample
/-- Example 1, p. 791: if `β` is the constant mean of the martingale
obtained by adding (3.8) and (3.9), then `β(1-T)=E(J₀+Y)>0`. -/
theorem example_1_beta {Ω : Type*} [mΩ : MeasurableSpace Ω]
(P : Measure Ω) [IsProbabilityMeasure P]
(𝓕 : Filtration ℝ≥0 mΩ) (T : ℝ≥0) (hT : 0 < T)
(J₀ Y : Ω → ℝ) (U V : ℝ≥0 → Ω → ℝ)
(husual : AntonelliBFSDE.BackwardForward.UsualHypotheses 𝓕 P)
(hJmeas : Measurable[𝓕 0] J₀) (hJint : Integrable J₀ P)
(hJpos : ∀ᵐ ω ∂P, 0 < J₀ ω)
(hYmeas : Measurable[𝓕 T] Y) (hYint : Integrable Y P)
(hYpos : ∀ᵐ ω ∂P, 0 < Y ω)
(hsol : IsDtSystemSolution 𝓕 P T J₀ (fun u v => u + v)
(fun u v => u + v) Y U V) :
(∀ᵐ r ∂((volume : Measure ℝ).restrict (Set.Ioc (0 : ℝ) (T : ℝ))),
Integrable (fun ω => U r.toNNReal ω + V r.toNNReal ω) P ∧
(∫ ω, U r.toNNReal ω + V r.toNNReal ω ∂P) =
(∫ ω, J₀ ω + xiDt T (fun u v => u + v) Y U V 0 ω ∂P) ∧
(∫ ω, U r.toNNReal ω + V r.toNNReal ω ∂P) =
(∫ ω, J₀ ω ∂P) +
(∫ s in Set.Ioc (0 : ℝ) (T : ℝ),
∫ ω, U s.toNNReal ω + V s.toNNReal ω ∂P) +
(∫ ω, Y ω ∂P)) ∧
(∫ ω, J₀ ω + xiDt T (fun u v => u + v) Y U V 0 ω ∂P) *
(1 - (T : ℝ)) = ∫ ω, J₀ ω + Y ω ∂P ∧
0 < ∫ ω, J₀ ω + Y ω ∂P := by sorry
end AntonelliBFSDE.SingularExample
Source
Antonelli, Backward-forward stochastic differential equations, Ann. Appl. Probab. 3(3) (1993), p. 791, Example 1, displays after (3.9), https://doi.org/10.1214/aoap/1177005363
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.