Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Section 2, equation (1) — hypograph reformulation

Proved
WangMinimaxRegret.Reformulation.eq_1

by Shuze Chen · Oct 29, 2026 · Mathlib 0df444a (Lean v4.33.1)

linear-optimizationmechanism-designoperations-researchp2o-gran-per-chapterp2o-plan-paperp2o-v1

Let the nonnegative valuation interval and finite moment information be consistent, and let F\mathcal FF be the printed ambiguity class (including F(a)=0F(a)=0F(a)=0). Let M\mathcal MM be the normalized Stieltjes payment class, with its cumulative allocation budget. For a payment m∈Mm\in\mathcal Mm∈M and any real rrr, including negative rrr, the hypograph constraints satisfy

[∀p∈I, rp≤inf⁡F∈F, TF(p)>0R(m,F)/TF(p)] ⟺ r≤C(m).\bigl[\forall p\in I,\ rp\le\inf_{F\in\mathcal F,\ T_F(p)>0}R(m,F)/T_F(p)\bigr]\ \Longleftrightarrow\ r\le C(m).[∀p∈I, rp≤F∈F, TF​(p)>0inf​R(m,F)/TF​(p)] ⟺ r≤C(m).

Consequently the value of (1) equals the original robust-screening value sup⁡m∈MC(m)\sup_{m\in\mathcal M}C(m)supm∈M​C(m). Here TF(p)=F([p,∞))T_F(p)=F([p,\infty))TF​(p)=F([p,∞)), R(m,F)=∫m dFR(m,F)=\int m\,dFR(m,F)=∫mdF, and C(m)C(m)C(m) is the supremum of nonnegative rates that satisfy all original price-wise revenue inequalities. This is the hypograph reformulation of (P).

The statement also records the model agreement used by the page: if every F∈FF\in\mathcal FF∈F has a finite strictly positive benchmark B(F)=sup⁡p∈IpTF(p)B(F)=\sup_{p\in I}pT_F(p)B(F)=supp∈I​pTF​(p), then

C(m)=inf⁡F∈FR(m,F)/B(F).C(m)=\inf_{F\in\mathcal F}R(m,F)/B(F).C(m)=F∈Finf​R(m,F)/B(F).

For any finite upper bound bbb and any F∈FF\in\mathcal FF∈F, the benchmark is attained by some price in III, as the displayed maximum requires. The finite-benchmark condition applies only to the quotient identification; it does not restrict the hypograph or value-equality clauses.

Differences from the source:

  • encoding: real interval, payment functions, probability laws, and moment orders 1 through K → real-line Borel measures supported on the interval, Stieltjes functions extended by zero below a and constantly above finite b, and Fin K exponents i+1; reason: extensions only fix values outside the domain; all atoms and payment jumps are retained, including the payment atom at a>0
  • encoding: the seller knows moments of an actual valuation distribution → Consistent d requires a nonempty printed ambiguity class, with finite specified moments; reason: makes the standing information-consistency assumption explicit; no interiority, positive variance, or density is assumed
  • reading: the notation [a,b] also occurs with b=infinity later in the paper → b has type WithTop real; prices and valuations remain finite; hindsight revenue is a supremum, and the competitive ratio uses all price-wise revenue inequalities; reason: specifies the support-only unbounded cases where the printed ratio may be infinity/infinity; finite positive benchmarks retain the usual quotient
  • encoding: nonnegative revenue integrals and optimization infima/suprema → nonnegative extended integrals and ENNReal/EReal extrema; zero-tail laws impose no fractional constraint; reason: avoids nonintegrable real integrals defaulting to zero and empty real extrema; empty infima are positive infinity
Preamble
import Definitions.Def_WangMinimaxRegret_Reformulation_Revenue

open MeasureTheory Set
open scoped ENNReal NNReal
set_option autoImplicit false
Formal statement
namespace WangMinimaxRegret.Reformulation

theorem eq_1 (d : Data) (hc : Consistent d) :
    (∀ (m : StieltjesFunction ℝ) (r : ℝ), IsPayment d m →
      (HypographFeasible d m r ↔ (r : EReal) ≤ (competitiveRatio d m : EReal))) ∧
    hypographValue d = originalValue d ∧
    (∀ m : StieltjesFunction ℝ, IsPayment d m →
      (∀ F : Measure ℝ, Ambiguous d F → 0 < benchmark d F ∧ benchmark d F < ⊤) →
      competitiveRatio d m = quotientRatio d m) ∧
    (∀ (b : ℝ) (F : Measure ℝ), d.b = (b : WithTop ℝ) → Ambiguous d F →
      ∃ p ∈ interval d, benchmark d F = priceRevenue F p) := by sorry

end WangMinimaxRegret.Reformulation
Source
Wang, Liu and Zhang (2024), Minimax Regret Robust Screening with Moment Information, https://doi.org/10.1287/msom.2023.0072, Section 2, p. 8, equation (1) and the immediately preceding ratio displays
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Setting. Fix a data record ddd:

  • reals a≥0a \ge 0a≥0 and b∈R∪{+∞}b \in \mathbb R\cup\{+\infty\}b∈R∪{+∞} with a≤ba \le ba≤b;
  • K∈NK \in \mathbb NK∈N and reals μ0,…,μK−1\mu_0,\dots,\mu_{K-1}μ0​,…,μK−1​;
  • the interval I={x:a≤x≤b}I = \{x : a \le x \le b\}I={x:a≤x≤b}.

A Borel measure FFF on R\mathbb RR is ambiguous when:

  • FFF is a probability measure;
  • F(R∖I)=0F(\mathbb R \setminus I) = 0F(R∖I)=0;
  • xi+1x^{i+1}xi+1 is FFF-integrable with ∫xi+1 dF=μi\int x^{i+1}\,dF = \mu_i∫xi+1dF=μi​ for all i<Ki < Ki<K;
  • F((−∞,a])=0F((-\infty,a]) = 0F((−∞,a])=0.

Hypothesis. ddd is consistent: at least one ambiguous measure exists.

Quantities used. Here mmm ranges over Stieltjes functions (nondecreasing, right-continuous, real-valued), and quantities are taken in [0,∞][0,\infty][0,∞] or R‾\overline{\mathbb R}R.

  • Payment. mmm is a payment when:
    • m(0)=0m(0)=0m(0)=0;
    • m=0m = 0m=0 on (−∞,a)(-\infty,a)(−∞,a);
    • m(x)=m(y)m(x) = m(y)m(x)=m(y) whenever x,y≥bx,y \ge bx,y≥b;
    • m≥0m \ge 0m≥0 on III;
    • ∫I∩(0,x]t−1 dνm(t)≤1\int_{I\cap(0,x]} t^{-1}\,d\nu_m(t) \le 1∫I∩(0,x]​t−1dνm​(t)≤1 for x∈Ix\in Ix∈I, with νm\nu_mνm​ the Lebesgue–Stieltjes measure of mmm.
  • Revenue. Rev(m,F)=∫max⁡(m,0) dF\mathrm{Rev}(m,F) = \int \max(m,0)\,dFRev(m,F)=∫max(m,0)dF.
  • Price revenue. PR(F,p)=max⁡(p,0) F([p,∞))\mathrm{PR}(F,p) = \max(p,0)\,F([p,\infty))PR(F,p)=max(p,0)F([p,∞)).
  • Benchmark. Bench(F)=sup⁡p∈IPR(F,p)\mathrm{Bench}(F) = \sup_{p\in I}\mathrm{PR}(F,p)Bench(F)=supp∈I​PR(F,p).
  • Competitive ratio. CR(m)\mathrm{CR}(m)CR(m) is the supremum of reals r≥0r \ge 0r≥0 with r PR(F,p)≤Rev(m,F)r\,\mathrm{PR}(F,p) \le \mathrm{Rev}(m,F)rPR(F,p)≤Rev(m,F) for all ambiguous FFF and all p∈Ip\in Ip∈I.
  • Quotient ratio. QR(m)=inf⁡F ambiguousRev(m,F)/Bench(F)\mathrm{QR}(m) = \inf_{F \text{ ambiguous}} \mathrm{Rev}(m,F)/\mathrm{Bench}(F)QR(m)=infF ambiguous​Rev(m,F)/Bench(F).
  • Hypograph feasibility. (m,r)(m,r)(m,r) is hypograph-feasible when mmm is a payment and, for every p∈Ip\in Ip∈I,
r p≤inf⁡{Rev(m,F)F([p,∞)):F ambiguous, F([p,∞))>0}.r\,p \le \inf\Big\{\tfrac{\mathrm{Rev}(m,F)}{F([p,\infty))} : F \text{ ambiguous},\ F([p,\infty))>0\Big\}.rp≤inf{F([p,∞))Rev(m,F)​:F ambiguous, F([p,∞))>0}.
  • Hypograph value. HV(d)=sup⁡{r:(m,r) hypograph-feasible}\mathrm{HV}(d) = \sup\{r : (m,r) \text{ hypograph-feasible}\}HV(d)=sup{r:(m,r) hypograph-feasible}.
  • Original value. OV(d)=sup⁡{CR(m):m a payment}\mathrm{OV}(d) = \sup\{\mathrm{CR}(m) : m \text{ a payment}\}OV(d)=sup{CR(m):m a payment}.

Conclusions. Under the hypothesis, all four of the following hold.

  1. Feasibility versus ratio. For every Stieltjes mmm that is a payment and every real rrr,
(m,r) is hypograph-feasible  ⟺  r≤CR(m)in R‾.(m,r) \text{ is hypograph-feasible} \iff r \le \mathrm{CR}(m) \quad\text{in } \overline{\mathbb R}.(m,r) is hypograph-feasible⟺r≤CR(m)in R.
  1. Equal values. HV(d)=OV(d)\mathrm{HV}(d) = \mathrm{OV}(d)HV(d)=OV(d) in R‾\overline{\mathbb R}R.
  2. Ratio as a quotient. For every payment mmm: if every ambiguous FFF satisfies 0<Bench(F)<∞0 < \mathrm{Bench}(F) < \infty0<Bench(F)<∞, then CR(m)=QR(m)\mathrm{CR}(m) = \mathrm{QR}(m)CR(m)=QR(m).
  3. Benchmark attained. For every real b′b'b′ and every measure FFF: if the upper endpoint is the finite real b′b'b′ (so b=b′b = b'b=b′) and FFF is ambiguous, then there is a p∈Ip \in Ip∈I with Bench(F)=PR(F,p)\mathrm{Bench}(F) = \mathrm{PR}(F,p)Bench(F)=PR(F,p).

Degenerate cases.

  • b=ab = ab=a. Consistency excludes it, since I={a}I=\{a\}I={a} cannot carry an ambiguous measure. So III has more than one point.
  • K=0K = 0K=0. Ambiguous measures are all probability measures on III with no mass at or below aaa.
  • Negative rrr. In conclusion 1, every r≤0r \le 0r≤0 lies below CR(m)≥0\mathrm{CR}(m) \ge 0CR(m)≥0. Since p≥0p \ge 0p≥0 on III, the equivalence then says such pairs are feasible.
  • Positive benchmark. Every ambiguous FFF has all its mass in (a,b](a,b](a,b] with a≥0a \ge 0a≥0, so its benchmark is automatically positive. The hypothesis in conclusion 3 therefore restricts only finiteness.
  • Finite benchmark. When bbb is finite, the benchmark is automatically at most bbb, so finiteness holds as well.
  • Infinite bbb. Conclusion 4 says nothing when b=+∞b = +\inftyb=+∞.
  • Revenue and division conventions. Revenues may be +∞+\infty+∞, and max⁡(m,0)\max(m,0)max(m,0) replaces negative values of mmm (the payment conditions already give m≥0m \ge 0m≥0 on III). Division follows [0,∞][0,\infty][0,∞] conventions (x/0=∞x/0 = \inftyx/0=∞ for x>0x>0x>0, 0/0=00/0 = 00/0=0). Zero-tail measures are skipped in the hypograph infimum, which is +∞+\infty+∞ when no ambiguous FFF has a positive tail at ppp.
Human review
  • Endorsed by Community (Bot) · Oct 29, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Shuze Chen · Oct 29, 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