Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 7.9 / Corollary 7.10: for a reward process with a location parameter and a conjugate prior with location parameter x̄, r(x̄ + c, n) = r(x̄, n) + c and ν(x̄, n) = x̄ + ν(0, n)

Proved
AllocationIndices.location_invariance_of_index

by naimengye · Sep 24, 2026 · Mathlib 0df444a (Lean v4.33.1)

conjugate-priorsgittins-indexinvariancereward-processes

Theorem 7.9 / Corollary 7.10. If the parameter μ\muμ for the family of distributions for a reward process is a location parameter, the sample mean Xˉ\bar XXˉ is a sufficient statistic, and μ\muμ has a conjugate prior distribution with the parameters xˉ\bar xxˉ and nnn, where xˉ\bar xxˉ is a location parameter and n≥0n \ge 0n≥0, and these parameters take the values (nxˉ+x)/(n+1)(n\bar x + x)/(n + 1)(nxˉ+x)/(n+1) and n+1n + 1n+1 when xxx is the next value sampled, then πn\pi_nπn​ may be identified by its parameters xˉ\bar xxˉ and nnn, and ν(xˉ,n)=xˉ+ν(0,n)\nu(\bar x, n) = \bar x + \nu(0, n)ν(xˉ,n)=xˉ+ν(0,n).

Formally: for a sampling model on Θ=R\Theta = \mathbb{R}Θ=R with parameters (xˉ,n)(\bar x, n)(xˉ,n), conjugate for n>0n > 0n>0, likelihood with location parameter μ\muμ (HasLocationParameter), prior family with location parameter xˉ\bar xxˉ (PriorHasLocationParameter), update meanUpdate, integrable observations (the mean of f(⋅∣θ)f(\cdot \mid \theta)f(⋅∣θ) exists, p. 206), a∈(0,1)a \in (0, 1)a∈(0,1), and the discounted rewards of the chain integrable from every state (L&S Assumption 35.6): r(xˉ+c,n)=r(xˉ,n)+cr(\bar x + c, n) = r(\bar x, n) + cr(xˉ+c,n)=r(xˉ,n)+c for every ccc, and the Gittins index of the reward process in the state (xˉ,n)(\bar x, n)(xˉ,n) equals xˉ+ν(0,n)\bar x + \nu(0, n)xˉ+ν(0,n). The book's improper uniform prior is replaced, as in Corollary 7.10, by a proper conjugate family with a location parameter; the identity R(λ+c,xˉ+c,n)=c(1−a)−1+R(λ,xˉ,n)R(\lambda + c, \bar x + c, n) = c(1-a)^{-1} + R(\lambda, \bar x, n)R(λ+c,xˉ+c,n)=c(1−a)−1+R(λ,xˉ,n) for the optimal payoff of {S,Λ}\{S, \Lambda\}{S,Λ} is not restated.

Admissible parameters. Conjugacy is required, and the conclusion stated, for n>0n > 0n>0 only (IsConjugateOn {p | 0 < p.2}, 0 < n): a proper prior has n>0n > 0n>0, the book's n=0n = 0n=0 being the improper prior of the theorem itself. Conjugacy at every (xˉ,n)∈R2(\bar x, n) \in \mathbb{R}^2(xˉ,n)∈R2 cannot hold together with a location parameter. At n=−1n = -1n=−1 the update (nxˉ+x)/(n+1)(n\bar x + x)/(n+1)(nxˉ+x)/(n+1) divides by zero (Lean's value is 000), so every observation leads to the same state (0,0)(0, 0)(0,0). Conjugacy would then force the posterior of the prior (xˉ,−1)(\bar x, -1)(xˉ,−1) to be the prior (0,0)(0, 0)(0,0) for every xˉ\bar xxˉ, but by the location structure that posterior moves with xˉ\bar xxˉ. Stated with conjugacy over all of R2\mathbb{R}^2R2, the theorem had no instances.

Preamble
import Definitions.Def_AllocationIndices_Sampling

open MeasureTheory ProbabilityTheory BanditAlgorithm
Formal statement
namespace AllocationIndices

/-- **Theorem 7.9 / Corollary 7.10** (pp. 189–191). If the parameter `μ` of the distributions of a
reward process is a location parameter, the sample mean is a sufficient statistic, and `μ` has a
conjugate prior distribution with parameters `x̄` and `n`, where `x̄` is a location 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 the state is identified by `(x̄, n)`, `r(x̄ + c, n) = r(x̄, n) + c`, and
`ν(x̄, n) = x̄ + ν(0, n)`. Stated for the Gittins index of the chain of parameters, under the
book's standing assumptions that the observations have a mean and the discounted rewards are
integrable (L&S Assumption 35.6). -/
theorem location_invariance_of_index (F : SamplingModel ℝ (ℝ × ℝ))
    (hconj : F.IsConjugateOn {p | 0 < p.2})
    (hlik : HasLocationParameter F.likelihood) (hprior : PriorHasLocationParameter 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) :
    (∀ c : ℝ, F.rewardOf (xb + c, n) = F.rewardOf (xb, n) + c) ∧
    gittinsIndex F.rewardChain F.rewardOf a (xb, n) =
      xb + gittinsIndex F.rewardChain F.rewardOf a (0, n) := by sorry

end AllocationIndices
Source
Gittins, Glazebrook and Weber, Multi-armed Bandit Allocation Indices, 2nd ed., Wiley 2011, doi:10.1002/9780470980033, §7.4 pp. 189-191, Theorem 7.9 with its proof ((7.17)-(7.21)) and Corollary 7.10
Human review
  • Endorsed by Shuze Chen · Sep 25, 2026

    Confirmed by the moderator at approval.

  • Endorsed by naimengye · Sep 25, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me