OAI/Probability/SLE/Geometry/RestartContinuationKernel.lean
11 declarations, 0 contain sorry, 0 declared as axioms
Conditional path tests after a finite restart. Splitting again at τ+t identifies the conditional mean in the full original stopped filtration; no factorization of its events through the restarted path is required.
- def
OAI.SLEExactGauge.driverContinuationMeanline 19 - lemma
OAI.SLEExactGauge.driverContinuation_test_measurableline 23 - lemma
OAI.SLEExactGauge.driverContinuationMean_measurableline 32 - lemma
OAI.SLEExactGauge.sourceRestartDriverPath_measurableline 38 - theorem
OAI.SLEExactGauge.source_driver_continuation_condLExpline 48 - theorem
OAI.SLEExactGauge.driverContinuationMean_eq_of_past_integralsline 82Source documentation
Deterministic-time conditional identities identify the continuation section on the canonical driver law.
theorem driverContinuationMean_eq_of_past_integrals {Ω : Type*} [mΩ : MeasurableSpace Ω] {P : Measure Ω} [IsProbabilityMeasure P] {B : ℝ≥0 → Ω → ℝ} (hB : IsBrownianReal B P) (hm : ∀ t, Measurable (B t)) (hc : ∀ ω, Continuous (fun t => B t ω)) (t : ℝ≥0) (f K : C(ℝ≥0,ℝ) → ℝ≥0∞) (hf : Measurable f) (hK : Measurable[driverPathFiltration t] K) (hmean : ∀ A : Set Ω, MeasurableSet[(Filtration.natural B (fun s => (hm s).stronglyMeasurable)) t] A → (∫⁻ ω in A, f (sourceDriverPath B hc ω) ∂P)= ∫⁻ ω in A, K (sourceDriverPath B hc ω) ∂P) : driverContinuationMean (P.map (sourceDriverPath B hc)) t f =ᵐ[P.map (sourceDriverPath B hc)] K := by …Mentions: OAI.SLEExactGauge.driverPathFiltration, OAI.SLEExactGauge.sourceDriverPath, OAI.SLEExactGauge.driverContinuationMean
- theorem
OAI.SLEExactGauge.source_restart_continuation_condLExpline 108Source documentation
The conditioning sigma algebra is the full original past at τ+t.
theorem source_restart_continuation_condLExp {Ω : Type*} [mΩ : MeasurableSpace Ω] {P : Measure Ω} [IsProbabilityMeasure P] {B : ℝ≥0 → Ω → ℝ} (hB : IsBrownianReal B P) (hm : ∀ t, Measurable (B t)) (hc : ∀ ω, Continuous (fun t => B t ω)) (τ : Ω → ℝ≥0) (hτ : IsStoppingTime (Filtration.natural B (fun s => (hm s).stronglyMeasurable)) (fun ω => (τ ω : WithTop ℝ≥0))) (t : ℝ≥0) (f : C(ℝ≥0,ℝ) → ℝ≥0∞) (hf : Measurable f) : (fun ω => driverContinuationMean (P.map (sourceDriverPath B hc)) t f (sourceRestartDriverPath B hc τ ω)) =ᵐ[P] P⁻[(fun ω => f (sourceRestartDriverPath B hc τ ω)) | shiftedStoppingFiltration hτ t] := by …Mentions: OAI.SLEExactGauge.driverContinuationMean, OAI.SLEExactGauge.sourceDriverPath, OAI.SLEExactGauge.sourceRestartDriverPath, OAI.SLEExactGauge.shiftedStoppingFiltration
- theorem
OAI.SLEExactGauge.source_restart_condLExp_of_section_aeline 145Source documentation
A canonical section identity therefore holds in the larger shifted filtration, not merely in the natural filtration of the fresh driver.
theorem source_restart_condLExp_of_section_ae {Ω : Type*} [mΩ : MeasurableSpace Ω] {P : Measure Ω} [IsProbabilityMeasure P] {B : ℝ≥0 → Ω → ℝ} (hB : IsBrownianReal B P) (hm : ∀ t, Measurable (B t)) (hc : ∀ ω, Continuous (fun t => B t ω)) (τ : Ω → ℝ≥0) (hτ : IsStoppingTime (Filtration.natural B (fun s => (hm s).stronglyMeasurable)) (fun ω => (τ ω : WithTop ℝ≥0))) (t : ℝ≥0) (f K : C(ℝ≥0,ℝ) → ℝ≥0∞) (hf : Measurable f) (hK : Measurable K) (hsection : driverContinuationMean (P.map (sourceDriverPath B hc)) t f =ᵐ[P.map (sourceDriverPath B hc)] K) : (fun ω => K (sourceRestartDriverPath B hc τ ω)) =ᵐ[P] P⁻[(fun ω => f (sourceRestartDriverPath B hc τ ω)) | shiftedStoppingFiltration hτ t] := by …Mentions: OAI.SLEExactGauge.driverContinuationMean, OAI.SLEExactGauge.sourceDriverPath, OAI.SLEExactGauge.sourceRestartDriverPath, OAI.SLEExactGauge.shiftedStoppingFiltration
- theorem
OAI.SLEExactGauge.source_restart_integral_of_section_aeline 169 - theorem
OAI.SLEExactGauge.source_restart_integral_of_original_kernelline 190 - theorem
OAI.SLEExactGauge.source_restart_bind_of_original_kernelline 213Source documentation
Measure-kernel form of the continuation identity. This applies to the sum of the finite-time raw mass and its current Green measure.
theorem source_restart_bind_of_original_kernel {Ω : Type*} [mΩ : MeasurableSpace Ω] {P : Measure Ω} [IsProbabilityMeasure P] {B : ℝ≥0 → Ω → ℝ} (hB : IsBrownianReal B P) (hm : ∀ t, Measurable (B t)) (hc : ∀ ω, Continuous (fun t => B t ω)) (t : ℝ≥0) (N K : C(ℝ≥0,ℝ) → Measure upperHalfPlane) (hN : Measurable N) (hK : Measurable[driverPathFiltration t] K) (hmean : ∀ A : Set Ω, MeasurableSet[(Filtration.natural B (fun s => (hm s).stronglyMeasurable)) t] A → (P.restrict A).bind (fun ω => N (sourceDriverPath B hc ω))= (P.restrict A).bind (fun ω => K (sourceDriverPath B hc ω))) (τ : Ω → ℝ≥0) (hτ : IsStoppingTime (Filtration.natural B (fun s => (hm s).stronglyMeasurable)) …Mentions: OAI.SLEExactGauge.upperHalfPlane, OAI.SLEExactGauge.driverPathFiltration, OAI.SLEExactGauge.sourceDriverPath