Theorem 7.11 / Corollary 7.12: for a reward process with a scale parameter and a conjugate prior with scale parameter x̄, ν(x̄, n) = x̄ ν(1, n)
ProvedAllocationIndices.scale_invariance_of_indexTheorem 7.11 / Corollary 7.12. If the parameter for the family of distributions for a reward process is a scale parameter, the sample mean is a sufficient statistic, and has a conjugate prior distribution with the parameters and , where is a scale parameter and the parameters take the values and when is the next value sampled, then may be identified by and and .
Formally: for a sampling model on with parameters , conjugate for , likelihood with scale parameter (HasScaleParameter), prior family with scale parameter (PriorHasScaleParameter), update meanUpdate, integrable observations, , and the discounted rewards of the chain integrable (L&S Assumption 35.6): for , and for the Gittins index of the reward process satisfies . (The book's first conclusion, for the optimal payoff of , is not restated.)
Admissible parameters. Conjugacy is required, and the conclusion stated, for only (IsConjugateOn {p | 0 < p.2}, 0 < n): a proper prior has , the book's being the improper prior of the theorem itself. Conjugacy at every cannot hold together with a location parameter. At the update divides by zero (Lean's value is ), so every observation leads to the same state . Conjugacy would then force the posterior of the prior to be the prior for every , but by the location structure that posterior moves with . Stated with conjugacy over all of , the theorem had no instances.
import Definitions.Def_AllocationIndices_Sampling open MeasureTheory ProbabilityTheory BanditAlgorithm
namespace AllocationIndices
/-- **Theorem 7.11 / Corollary 7.12** (pp. 191–192). If the parameter `σ > 0` of the distributions
of a reward process is a scale parameter, the sample mean is a sufficient statistic, and `σ` has a
conjugate prior distribution with parameters `x̄` and `n`, where `x̄` is a scale parameter of the
prior family and the parameters become `(n x̄ + x)/(n + 1)` and `n + 1` when `x` is the next
value sampled, then `r(b x̄, n) = b r(x̄, n)` for `b > 0` and `ν(x̄, n) = x̄ ν(1, n)` for `x̄ > 0`. -/
theorem scale_invariance_of_index (F : SamplingModel ℝ (ℝ × ℝ))
(hconj : F.IsConjugateOn {p | 0 < p.2})
(hlik : HasScaleParameter F.likelihood) (hprior : PriorHasScaleParameter F.prior)
(hupd : F.update = meanUpdate) (hmean : ∀ p, Integrable (fun x : ℝ ↦ x) (F.predictive p))
{a : ℝ} (ha0 : 0 < a) (ha1 : a < 1)
(hint : DiscountedRewardIntegrable F.rewardChain F.rewardOf a) (xb n : ℝ) (hn : 0 < n) (hxb : 0 < xb) :
(∀ b : ℝ, 0 < b → F.rewardOf (b * xb, n) = b * F.rewardOf (xb, n)) ∧
gittinsIndex F.rewardChain F.rewardOf a (xb, n) =
xb * gittinsIndex F.rewardChain F.rewardOf a (1, n) := by sorry
end AllocationIndices
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.