Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 10.1 — under a Markov (PSI) policy, Vp,kV_{p,k}Vp,k​ carries Pk(π^,p)P_k(\hat\pi,p)Pk​(π^,p) to P^k[π^,φ(p)]\hat P_k[\hat\pi,\varphi(p)]P^k​[π^,φ(p)]

Disproved
BertsekasShreve.ImperfectInfo.lemma_10_1

by mikedeng1 · Oct 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

dynamic-programmingimperfect-informationp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-bookp2o-v1sufficient-statistic

Let an (ISI) model with horizon NNN and a statistic (η0,…,ηN−1)(\eta_0,\dots,\eta_{N-1})(η0​,…,ηN−1​) sufficient for control be given, with associated (PSI) model. Let p∈P(S)p\in P(S)p∈P(S) and let π^=(μ^0,μ^1,… )\hat\pi=(\hat\mu_0,\hat\mu_1,\dots)π^=(μ^​0​,μ^​1​,…) be a Markov (PSI) policy, used in (ISI) through μk(du∣p;ik)=μ^k(du∣ηk(p;ik))\mu_k(du\mid p;i_k)=\hat\mu_k(du\mid\eta_k(p;i_k))μk​(du∣p;ik​)=μ^​k​(du∣ηk​(p;ik​)). Then for k=0,…,N−1k=0,\dots,N-1k=0,…,N−1 and every Borel set B⊆Y0C0⋯YkCkB\subseteq Y_0C_0\cdots Y_kC_kB⊆Y0​C0​⋯Yk​Ck​,

Pk(π^,p)[Vp,k−1(B)]=P^k[π^,φ(p)](B),P_k(\hat\pi,p)\big[V_{p,k}^{-1}(B)\big]=\hat P_k[\hat\pi,\varphi(p)](B),Pk​(π^,p)[Vp,k−1​(B)]=P^k​[π^,φ(p)](B),

where Vp,k(x0,z0,u0,…,xk,zk,uk)=[η0(p;i0),u0,…,ηk(p;ik),uk]V_{p,k}(x_0,z_0,u_0,\dots,x_k,z_k,u_k)=[\eta_0(p;i_0),u_0,\dots,\eta_k(p;i_k),u_k]Vp,k​(x0​,z0​,u0​,…,xk​,zk​,uk​)=[η0​(p;i0​),u0​,…,ηk​(p;ik​),uk​] and φ(p)\varphi(p)φ(p) is the distribution (26) of y0=η0(p;z0)y_0=\eta_0(p;z_0)y0​=η0​(p;z0​).

In words: the statistic process (η0,u0,…,ηk,uk)(\eta_0,u_0,\dots,\eta_k,u_k)(η0​,u0​,…,ηk​,uk​) generated in (ISI) by π^\hat\piπ^ has exactly the law of the state–control process of (PSI) under the same policy started from φ(p)\varphi(p)φ(p). This is the measure-level identity behind Proposition 10.2.

Preamble
import Mathlib
import Definitions.Def_BertsekasShreve_ImperfectInfo_SufficientStatistic

open MeasureTheory ProbabilityTheory
Formal statement
namespace BertsekasShreve.ImperfectInfo

/-- Lemma 10.1 (p. 253). -/
theorem lemma_10_1 {S C Z : Type}
    [TopologicalSpace S] [MeasurableSpace S] [BorelSpace S]
    [TopologicalSpace C] [MeasurableSpace C] [BorelSpace C]
    [TopologicalSpace Z] [MeasurableSpace Z] [BorelSpace Z]
    (M : ISIModel S C Z) {Y : ℕ → Type} [∀ k, TopologicalSpace (Y k)] [∀ k, MeasurableSpace (Y k)]
    (σ : SuffStat M Y)
    (p : ProbabilityMeasure S) (μ : (k : ℕ) → Y k → ProbabilityMeasure C)
    (hμ : σ.IsMarkovPolicy μ) (k : ℕ) (hk : k < M.N) (B : Set (Hist Y C k))
    (hB : MeasurableSet B) :
    M.law (σ.toISI μ) p k (σ.V p k ⁻¹' B) = σ.lawHat (markovToPolicy μ) (σ.phi p) k B := by sorry

end BertsekasShreve.ImperfectInfo
Source
Bertsekas & Shreve, Stochastic Optimal Control: The Discrete-Time Case, Athena Scientific 1996, p. 253, Lemma 10.1 (Eq. (29) of Chapter 10)
Human review
  • Endorsed by Shuze Chen · Oct 5, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 5, 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