Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proof of Theorem 4.18, p. 31 — an increasing family SxS_xSx​ with Sx(R)=μ(]−∞,x])S_x(\mathbb R)=\mu(]-\infty,x])Sx​(R)=μ(]−∞,x]) defines a plan with π(]−∞,x]×A)=Sx(A)\pi(]-\infty,x]\times A)=S_x(A)π(]−∞,x]×A)=Sx​(A)

Proved
MartOT.Curtain.curtain_measure_exists

by tianyipeng · Nov 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

left-curtain-couplingmartingale-optimal-transportmeasure-construction

Let μ\muμ be a finite Borel measure on R\mathbb RR and let (Sx)x∈R(S_x)_{x\in\mathbb R}(Sx​)x∈R​ be a family of finite Borel measures on R\mathbb RR that is increasing, i.e. Sx≤Sx′S_x\le S_{x'}Sx​≤Sx′​ setwise whenever x≤x′x\le x'x≤x′, and whose masses are Sx(R)=μ(]−∞,x])S_x(\mathbb R)=\mu(]-\infty,x])Sx​(R)=μ(]−∞,x]) for every xxx. Then there is a finite measure π\piπ on R×R\mathbb R\times\mathbb RR×R such that, for every x∈Rx\in\mathbb Rx∈R,

proj#x(π∣]−∞,x]×R)=μ∣]−∞,x]andproj#y(π∣]−∞,x]×R)=Sx,\mathrm{proj}^x_\#\big(\pi|_{]-\infty,x]\times\mathbb R}\big)=\mu|_{]-\infty,x]}\qquad\text{and}\qquad \mathrm{proj}^y_\#\big(\pi|_{]-\infty,x]\times\mathbb R}\big)=S_x ,proj#x​(π∣]−∞,x]×R​)=μ∣]−∞,x]​andproj#y​(π∣]−∞,x]×R​)=Sx​,

that is, π(]−∞,x]×A)=Sx(A)\pi(]-\infty,x]\times A)=S_x(A)π(]−∞,x]×A)=Sx​(A) for all Borel AAA (the second condition is written with MartOT.Var.targetUpTo π x, the paper's νxπ\nu^\pi_xνxπ​).

This is the measure-construction step of the proof of Theorem 4.18 in Beiglböck–Juillet: "the condition given in the statement prescribes the value of πlc(]−∞,x]×A)=Sν(μ∣]−∞,x])(A)\pi_{lc}(]-\infty,x]\times A)=S^\nu(\mu|_{]-\infty,x]})(A)πlc​(]−∞,x]×A)=Sν(μ∣]−∞,x]​)(A) for x∈Rx\in\mathbb Rx∈R and every Borel set AAA, thus giving rise to a unique measure on the product space. Here we use that, by Theorem 4.8, Sν(μ∣]−∞,x])≤Sν(μ∣]−∞,x′])S^\nu(\mu|_{]-\infty,x]})\le S^\nu(\mu|_{]-\infty,x']})Sν(μ∣]−∞,x]​)≤Sν(μ∣]−∞,x′]​) whenever x≤x′x\le x'x≤x′." Here it is stated for an abstract increasing family. The mass condition gives that x↦Sx(A)x\mapsto S_x(A)x↦Sx​(A) is right-continuous and vanishes at −∞-\infty−∞: for x<x′x<x'x<x′ we have Sx′(A)−Sx(A)≤μ(]x,x′])S_{x'}(A)-S_x(A)\le\mu(]x,x'])Sx′​(A)−Sx​(A)≤μ(]x,x′]). So each x↦Sx(A)x\mapsto S_x(A)x↦Sx​(A) is the distribution function of a measure mA≤μm_A\le\mumA​≤μ on the xxx-axis, and π\piπ is obtained by disintegrating against μ\muμ (for instance π=μ⊗κ\pi=\mu\otimes\kappaπ=μ⊗κ with a kernel κ\kappaκ built from the densities dm]−∞,r]/dμdm_{]-\infty,r]}/d\mudm]−∞,r]​/dμ, r∈Qr\in\mathbb Qr∈Q).

Preamble
import Mathlib
import Definitions.Def_MartOT_Var_Setting
Formal statement
namespace MartOT.Curtain

open MeasureTheory

theorem curtain_measure_exists (μ : Measure ℝ) [IsFiniteMeasure μ] (S : ℝ → Measure ℝ)
    (hfin : ∀ x, IsFiniteMeasure (S x)) (hmono : ∀ x x', x ≤ x' → S x ≤ S x')
    (hmass : ∀ x, S x Set.univ = μ (Set.Iic x)) :
    ∃ π : Measure (ℝ × ℝ), IsFiniteMeasure π ∧
      ∀ x : ℝ, (π.restrict (Set.Iic x ×ˢ Set.univ)).map Prod.fst = μ.restrict (Set.Iic x) ∧
        MartOT.Var.targetUpTo π x = S x := by sorry

end MartOT.Curtain
Source
arXiv:1208.1509v2 (Beiglböck–Juillet, On a problem of optimal transport under marginal martingale constraints), proof of Theorem 4.18, p. 31 (construction of π_lc from the increasing family of shadows)

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