Extend eventwise conditional laws to test-integral identities
ProvedMeasureTheory.eventwiseLaw_integral_test_identitybanditsp2o-batch-p100ap2o-gran-per-chapterp2o-plan-paperp2o-v1revenue-managementthompson-sampling
Under the eventwise conditional-law hypotheses and the almost-sure bound, the realized observable and its kernel mean are integrable; the kernel mean is measurable with respect to the conditioning sigma algebra almost everywhere; and their integrals agree on every conditioning-measurable set.
Preamble
import Mathlib.MeasureTheory.Function.ConditionalExpectation.Basic open MeasureTheory open scoped MeasureTheory
Formal statement
theorem MeasureTheory.eventwiseLaw_integral_test_identity
{Ω : Type*} {E : Type*} [mΩ : MeasurableSpace Ω] [mE : MeasurableSpace E]
(P : Measure Ω) [IsProbabilityMeasure P]
(m : MeasurableSpace Ω) (hm : m ≤ mΩ)
(X : Ω → E) (hX : @Measurable Ω E mΩ mE X)
(ν : Ω → Measure E) (hν : ∀ ω, IsProbabilityMeasure (ν ω))
(hLaw : ∀ B : Set E, MeasurableSet B →
P[(X ⁻¹' B).indicator (fun _ => (1 : ℝ)) | m] =ᵐ[P]
fun ω => (ν ω B).toReal)
(f : E → ℝ) (hf : Measurable f)
(C : ℝ) (hC : 0 ≤ C)
(hbound : ∀ ω, ∀ᵐ x ∂(ν ω), ‖f x‖ ≤ C) :
Integrable (fun ω => f (X ω)) P ∧
Integrable (fun ω => ∫ x, f x ∂(ν ω)) P ∧
AEStronglyMeasurable[m] (fun ω => ∫ x, f x ∂(ν ω)) P ∧
∀ s : Set Ω, MeasurableSet[m] s →
∫ ω in s, (∫ x, f x ∂(ν ω)) ∂P =
∫ ω in s, f (X ω) ∂P := by sorrySource
Ferreira, Simchi-Levi, Wang, Online Network Revenue Management Using Thompson Sampling, Oper. Res. 66(6), 2018, p. 1594, Theorem 1