Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

(4), §3.2, p. 16 — Bellman equation V_k = min_{h∈[k]} g_k(h) + V_{k−h} for the optimal IBR partition

Proved
CorrClusterExp.IBR.bellman_4

by mikedeng1 · Oct 23, 2026 · Mathlib 0df444a (Lean v4.33.1)

block-randomizationdynamic-programmingexperimental-designp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Let 0<q≤120<q\le\tfrac120<q≤21​ and let w1≥w2≥⋯≥wn≥0w_1\ge w_2\ge\dots\ge w_n\ge 0w1​≥w2​≥⋯≥wn​≥0 with n≥1n\ge 1n≥1. For h∈[n]h\in[n]h∈[n] let gn(h)g_n(h)gn​(h) be the worst-case variance of a block made of the first hhh clusters under the block law of Proposition 3.1, and let Vn−hV_{n-h}Vn−h​ be the worst-case variance of the optimal IBR experiment on the remaining clusters h+1,…,nh+1,\dots,nh+1,…,n (V0=0V_0=0V0​=0). Then the worst-case variance Vn=VDPV_n=V^{\mathrm{DP}}Vn​=VDP of the optimal IBR experiment satisfies the Bellman equation

Vn=min⁡h∈[n] gn(h)+Vn−h.(4)V_n=\min_{h\in[n]}\ g_n(h)+V_{n-h}. \tag{4}Vn​=h∈[n]min​ gn​(h)+Vn−h​.(4)

Applied to the last kkk clusters, this is the recursion Vk=min⁡h∈[k]gk(h)+Vk−hV_k=\min_{h\in[k]}g_k(h)+V_{k-h}Vk​=minh∈[k]​gk​(h)+Vk−h​ of the paper, which computes the optimal partition in polynomial time.

Formalization Note The minimum is stated as two clauses: Vn≤gn(h)+Vn−hV_n\le g_n(h)+V_{n-h}Vn​≤gn​(h)+Vn−h​ for every h∈[n]h\in[n]h∈[n], and equality for some h∈[n]h\in[n]h∈[n]. The suffix sizes are fun i : Fin (n - h) => w ⟨h + i, _⟩.

Differences from the source:

  • encoding: VkV_kVk​ for the last kkk of the nnn sorted clusters → VDPV^{\mathrm{DP}}VDP of the suffix size vector, with the equation stated for every nnn and www; reason: the recursion at level kkk is the instance on the suffix, so every level is covered.
  • encoding: min⁡h∈[k]\min_{h\in[k]}minh∈[k]​ → "≤\le≤ for all h∈[1,n]h\in[1,n]h∈[1,n] and === for some"; reason: same statement for a finite nonempty index set.
Preamble
import Mathlib
import Definitions.Def_CorrClusterExp_IBR_Setting
Formal statement
namespace CorrClusterExp.IBR

open Matrix Finset

theorem bellman_4 (q : ℝ) (hq0 : 0 < q) (hq : q ≤ 1 / 2) (n : ℕ) (hn : 1 ≤ n)
    (w : Fin n → ℝ) (hw0 : ∀ i, 0 ≤ w i) (hw : Antitone w) :
    (∀ (h : ℕ) (h1 : 1 ≤ h) (h2 : h ≤ n),
      VDP q w ≤ worstCase (corr q (fracLaw q h)) (fun i : Fin h => w (Fin.castLE h2 i)) +
        VDP q (fun i : Fin (n - h) => w ⟨h + i.val, by have := i.isLt; omega⟩)) ∧
    ∃ (h : ℕ) (h1 : 1 ≤ h) (h2 : h ≤ n),
      VDP q w = worstCase (corr q (fracLaw q h)) (fun i : Fin h => w (Fin.castLE h2 i)) +
        VDP q (fun i : Fin (n - h) => w ⟨h + i.val, by have := i.isLt; omega⟩) := by sorry

end CorrClusterExp.IBR
Source
Candogan, Chen, Niazadeh, Correlated Cluster-Based Randomized Experiments: Robust Variance Minimization, SSRN 3852100 (version of March 22, 2023), p. 16, §3.2, equation (4)
Human review
  • Endorsed by Shuze Chen · Oct 23, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 23, 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