Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma EC.2: continuous regular approximation

Proved
MinBuying.lemmaEC2_continuous_smoothing

by mikeh · Oct 30, 2026 · Mathlib 0df444a (Lean v4.33.1)

min-buyingoperations-researchpricingprobability

For a positive finite triangular instance whose optimal uniform revenue is at most one, there exists a positive smallness threshold such that every sufficiently small 0<δ<10<\delta<10<δ<1 admits an independent continuous regular instance with the same revenue at the original cap-price vector and optimal uniform revenue at most 1/(1−δ)1/(1-\delta)1/(1−δ).

Preamble
import Definitions.Def_MinBuying_Triangular
import Definitions.Def_MinBuying_Regularity

open MeasureTheory

Formal statement
namespace MinBuying

/-- Lemma EC.2, PDF pp.45--47: for every sufficiently small positive
smoothing width, cap-pricing revenue is preserved exactly while the
optimal uniform revenue increases by at most the displayed factor. -/
theorem lemmaEC2_continuous_smoothing (n : ℕ) (hn : 0 < n)
    (I : TriangularInstance n) (hUP : UPRevenue (triangularJoint I) ≤ 1) :
    ∃ δ₀ : ℝ, 0 < δ₀ ∧ ∀ δ : ℝ, 0 < δ → δ < min 1 δ₀ →
      ∃ F : Fin n → Measure ℝ,
        RegularInstance F ∧
        Revenue (productLaw F) I.cap = Revenue (triangularJoint I) I.cap ∧
        UPRevenue (productLaw F) ≤ 1 / (1 - δ) := by
  sorry

end MinBuying
Source
Titing Cui, Michael L. Hamilton, Jourdain Lamperski, and Kasra Ganjizadeh Tari, Unit-Demand Pricing for Min-Buying Customers, September 27, 2026 manuscript, 47 PDF pages (main pp.1–27; companion ec1–ec20). https://doi.org/10.2139/ssrn.5178133 . PDF pp.45–47/ec18–ec20, Lemma EC.2.
Read-back

What the Lean code literally says, in plain math · gpt-6-astra

For real a,ca,ca,c, let La,cL_{a,c}La,c​ be the image of Lebesgue measure restricted to (0,1)(0,1)(0,1) under u↦min⁡{c,au/(1−u)}u\mapsto\min\{c,au/(1-u)\}u↦min{c,au/(1−u)}. For a measure λ\lambdaλ on a finite real coordinate space and a price vector ppp, let Mp(x)M_p(x)Mp​(x) be the minimum of the numbers pip_ipi​ over those indices satisfying pi≤xip_i\le x_ipi​≤xi​, with Mp(x)=0M_p(x)=0Mp​(x)=0 when there is no such index, and put Rλ(p)=∫Mp(x) dλ(x)R_\lambda(p)=\int M_p(x)\,d\lambda(x)Rλ​(p)=∫Mp​(x)dλ(x). Write Dλ=sup⁡{Rλ(p):pi≥0 for every i}D_\lambda=\sup\{R_\lambda(p):p_i\ge0\text{ for every }i\}Dλ​=sup{Rλ​(p):pi​≥0 for every i} and Uλ=sup⁡{Rλ(r1):r≥0}U_\lambda=\sup\{R_\lambda(r\mathbf1):r\ge0\}Uλ​=sup{Rλ​(r1):r≥0}, where the suprema are taken in R\mathbb RR and r1r\mathbf1r1 is the constant price vector. Write W(λ)W(\lambda)W(λ) for the conjunction that both sets defining DλD_\lambdaDλ​ and UλU_\lambdaUλ​ are bounded above, there is a vector ppp with every pi≥0p_i\ge0pi​≥0 and Rλ(p)=DλR_\lambda(p)=D_\lambdaRλ​(p)=Dλ​, there is an r≥0r\ge0r≥0 with Rλ(r1)=UλR_\lambda(r\mathbf1)=U_\lambdaRλ​(r1)=Uλ​, and Uλ>0U_\lambda>0Uλ​>0. For every integer n>0n>0n>0 and arrays (ai,ci)i<n(a_i,c_i)_{i<n}(ai​,ci​)i<n​ of positive real numbers, let μ=⨂i<nLai,ci\mu=\bigotimes_{i<n}L_{a_i,c_i}μ=⨂i<n​Lai​,ci​​ and assume Uμ≤1U_\mu\le1Uμ​≤1. There exists a real δ0>0\delta_0>0δ0​>0 such that for every real δ\deltaδ with 0<δ<min⁡{1,δ0}0<\delta<\min\{1,\delta_0\}0<δ<min{1,δ0​} there exists a family (Fi)i<n(F_i)_{i<n}(Fi​)i<n​ of measures on R\mathbb RR with the following simultaneous properties. Every FiF_iFi​ is a probability measure concentrated on [0,∞)[0,\infty)[0,∞), is absolutely continuous with respect to Lebesgue measure, and satisfies the concavity condition described next; and W(ν)W(\nu)W(ν) holds for ν=⨂i<nFi\nu=\bigotimes_{i<n}F_iν=⨂i<n​Fi​. For each marginal GGG, the concavity condition means that the function q↦qQG(1−q)q\mapsto qQ_G(1-q)q↦qQG​(1−q) is concave on the open interval (0,1)(0,1)(0,1), where QG(u)=inf⁡{x∈R:0≤x, u≤G((−∞,x])}Q_G(u)=\inf\{x\in\mathbb R:0\le x,\ u\le G((-\infty,x])\}QG​(u)=inf{x∈R:0≤x, u≤G((−∞,x])} is the real infimum; the marginal masses in this formula are finite real numbers. The same family also satisfies Rν(c)=Rμ(c)R_\nu(c)=R_\mu(c)Rν​(c)=Rμ​(c), where c=(ci)i<nc=(c_i)_{i<n}c=(ci​)i<n​ is exactly the original cap-price vector, and Uν≤1/(1−δ)U_\nu\le1/(1-\delta)Uν​≤1/(1−δ). The parameter δ0\delta_0δ0​ is chosen before the universal choice of δ\deltaδ, whereas the family FFF may depend on δ\deltaδ. Both endpoint inequalities for δ\deltaδ are strict, so 1−δ>01-\delta>01−δ>0; the revenue equality is exact at the specified vector, and n=0n=0n=0 is excluded.

Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikeh · Oct 31, 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