Posterior sampling and demand laws induce a bounded finite oracle experiment
ProvedOnlineNRM.TSFixed.posterior_matching_finite_oracle_experimentbanditsp2o-batch-p100ap2o-gran-per-chapterp2o-plan-paperp2o-v1revenue-managementthompson-sampling
The posterior-sampling, randomized-offer, and bounded conditional-demand assumptions imply the abstract finite oracle experiment fields. Prove these using conditional-expectation towers, bounded-function extensions of the eventwise laws, kernel-integral measurability, and support bounds. The shut-off action has zero reward and unused selector mass. All hypotheses remain valid at t = 0, K = 1, and N = 0.
Preamble
import Definitions.Def_OnlineNRM_TSFixed_EntropyExperiment open MeasureTheory ProbabilityTheory OnlineNRM.TSFixed open scoped BigOperators OnlineNRM.TSFixed
Formal statement
theorem OnlineNRM.TSFixed.posterior_matching_finite_oracle_experiment
{N K : ℕ} (hK : 1 ≤ K)
(p : Fin N → Fin K → ℝ) (dbar : Fin N → ℝ)
(hp : ∀ i k, 0 ≤ p i k)
{Θ : Type*} [MeasurableSpace Θ] [StandardBorelSpace Θ]
(F : Kernel (Θ × Fin K) (Fin N → ℝ)) [IsMarkovKernel F]
(hF : ∀ ρ k, ∀ᵐ x ∂(F (ρ, k)), ∀ i, x i ∈ Set.Icc 0 (dbar i))
{Ω : Type*} [MeasurableSpace Ω]
(P : Measure Ω) [IsProbabilityMeasure P]
(θ : Ω → Θ) (θs : ℕ → Ω → Θ)
(A : ℕ → Ω → Option (Fin K)) (D : ℕ → Ω → Fin N → ℝ)
(xsel : Θ → Fin K → ℝ) (t : ℕ)
(hθ : Measurable θ) (hθs : ∀ s, Measurable (θs s))
(hA : ∀ s, Measurable (A s)) (hD : ∀ s, Measurable (D s))
(hx : Measurable xsel)
(hx0 : ∀ ρ k, 0 ≤ xsel ρ k)
(hx1 : ∀ ρ, ∑ k : Fin K, xsel ρ k ≤ 1)
(hsample : ∀ C : Set Θ, MeasurableSet C →
condProb P (pastSigma θ θs A D t) (θs t ⁻¹' C) =ᵐ[P]
condProb P (histSigma A D t) (θ ⁻¹' C))
(hoffer : ∀ k : Fin K,
condProb P (pastSigma θ θs A D t ⊔
MeasurableSpace.comap (θs t) inferInstance) (A t ⁻¹' {some k}) =ᵐ[P]
fun ω => xsel (θs t ω) k)
(hdemand : ∀ B : Set (Fin N → ℝ), MeasurableSet B →
condProb P (pastSigma θ θs A D t ⊔
MeasurableSpace.comap (θs t) inferInstance ⊔
MeasurableSpace.comap (A t) inferInstance) (D t ⁻¹' B) =ᵐ[P]
fun ω => (demandLaw F (θ ω) (A t ω) B).toReal) :
IsFiniteOracleExperiment P (histSigma A D t) (pastSigma θ θs A D t)
(histSigma A D (t + 1))
(fun ω a => oracleActionWeights xsel (θ ω) a)
(fun ω a => oracleMeanReward p F (θ ω) a)
(A t) (observedPriceReward p (A t) (D t)) (pmax p dbar) := by sorrySource
Ferreira, Simchi-Levi, Wang, Online Network Revenue Management Using Thompson Sampling, Oper. Res. 66(6), 2018, p. 1594, Theorem 1