Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 1, (7), p. 12 — E[(1/τ)Φ_τ(D^{τ,λ}, S^τ)] → Φ(λ, π) as τ → ∞

Proved
MarketInterference.FluidLimit.theorem_1

by mikedeng1 · Oct 14, 2026 · Mathlib 0df444a (Lean v4.33.1)

fluid-limitlinear-programmingmatchingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1poisson-distribution

Consider a matching network with values vu,wv_{u,w}vu,w​ and capacities ku,w≥0k_{u,w}\ge 0ku,w​≥0, demand rates λ∈R≥0nd\lambda\in\mathbb R^{n_d}_{\ge 0}λ∈R≥0nd​​ and supply rates π∈R≥0ns\pi\in\mathbb R^{n_s}_{\ge 0}π∈R≥0ns​​. For each integer density τ\tauτ let Diτ,λ∼Poisson(λiτ)D^{\tau,\lambda}_i\sim\mathrm{Poisson}(\lambda_i\tau)Diτ,λ​∼Poisson(λi​τ) and Sjτ∼Poisson(πjτ)S^\tau_j\sim\mathrm{Poisson}(\pi_j\tau)Sjτ​∼Poisson(πj​τ) be independent counts, and let Φτ\Phi_\tauΦτ​ be the optimal value of the matching program (2) with capacities τk\tau kτk. Then

lim⁡τ→∞E[1τ Φτ(Dτ,λ,Sτ)]=Φ(λ,π).\lim_{\tau\to\infty}\mathbb E\Big[\frac1\tau\,\Phi_\tau\big(D^{\tau,\lambda},S^\tau\big)\Big]=\Phi(\lambda,\pi).τ→∞lim​E[τ1​Φτ​(Dτ,λ,Sτ)]=Φ(λ,π).

In words: in a dense market the expected matching value per unit of density is the value of the deterministic linear program whose demand and supply are the arrival rates. This is what lets the rest of the paper study treatment effects through the fluid program Φ(λ,π)\Phi(\lambda,\pi)Φ(λ,π) instead of the stochastic system.

Formalization Note The limit is taken along integer τ→∞\tau\to\inftyτ→∞ ("we only consider integer densities τ", p. 12); at τ=0\tau=0τ=0 Lean's 1/0=01/0=01/0=0 gives a junk term that does not affect the limit. The expectation is the Bochner integral against the product of the independent Poisson laws; the integrand is bounded (0≤1τΦτ≤∑∣vu,w∣ku,w0\le\frac1\tau\Phi_\tau\le\sum|v_{u,w}|k_{u,w}0≤τ1​Φτ​≤∑∣vu,w​∣ku,w​) and every function on the countable count space is measurable, so it is integrable. Only the first sentence of Theorem 1, display (7), is stated; the second sentence (almost-sure convergence of optimal primal and dual solutions) needs unique optima, assumed only inside the proof, and a coupling of the counts across τ\tauτ that the paper does not construct.


Retired 2026-10-27: a Lean-vs-paper fidelity audit found this statement weaker than, or narrower than, the cited result. A corrected statement will be resubmitted.

Preamble
import Mathlib
import Definitions.Def_MarketInterference_FluidBias_NetworkLP
import Definitions.Def_MarketInterference_FluidLimit_PoissonCounts
Formal statement
namespace MarketInterference.FluidLimit

open MeasureTheory Filter Topology

/-- Theorem 1, display (7) (arXiv:2205.02274v4, p. 12): along integer densities `τ → ∞`,
`E[(1/τ) Φ_τ(D^{τ,λ}, S^τ)] → Φ(λ, π)`, where `D^{τ,λ}_i ∼ Poisson(λ_i τ)` and `S^τ_j ∼ Poisson(π_j τ)`
are independent. -/
theorem theorem_1 {nd ns : ℕ} {I E : Type} [Fintype E] [DecidableEq I]
    (N : MarketInterference.FluidBias.Network nd ns I E) (lam : Fin nd → ℝ) (pi : Fin ns → ℝ) (hlam : 0 ≤ lam) (hpi : 0 ≤ pi) :
    Tendsto (fun τ : ℕ => ∫ z, (1 / (τ : ℝ)) * MarketInterference.FluidBias.Phi N τ (fun i => (z.1 i : ℝ)) (fun j => (z.2 j : ℝ))
        ∂(fluidLaw lam pi τ)) atTop (𝓝 (MarketInterference.FluidBias.Phi N 1 lam pi)) := by sorry

end MarketInterference.FluidLimit
Source
Bright, Delarue and Lobel, Reducing Marketplace Interference Bias Via Shadow Prices, arXiv:2205.02274v4, p. 12, Theorem 1, (7)
Human review
  • Endorsed by Shuze Chen · Oct 14, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 14, 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