Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 1, p. 603 — X IGFR ⇔ log X IFR ⇔ X ⪯hr λX (λ ≥ 1) ⇔ Φ̄(ξ/θ) TP₂

Proved
LariviereIGFR.Char.theorem_1

by mikedeng1 · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

characterizationhazard-rate-orderifrigfrp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1total-positivity

Let XXX be a nonnegative random variable with distribution function Φ\PhiΦ, survival function Φˉ=1−Φ\bar\Phi = 1 - \PhiΦˉ=1−Φ and density φ\varphiφ, taken as the version that is the derivative of Φ\PhiΦ on the support (α,β)(\alpha,\beta)(α,β) and 000 to its left. Write h=φ/Φˉh = \varphi/\bar\Phih=φ/Φˉ for its failure rate and g(ξ)=ξh(ξ)g(\xi) = \xi h(\xi)g(ξ)=ξh(ξ) for its generalized failure rate. Let XL=log⁡XX_L = \log XXL​=logX, with density φL(ξ)=eξφ(eξ)\varphi_L(\xi) = e^\xi\varphi(e^\xi)φL​(ξ)=eξφ(eξ), and for λ≥1\lambda \ge 1λ≥1 let λX\lambda XλX have density φ(ξ/λ)/λ\varphi(\xi/\lambda)/\lambdaφ(ξ/λ)/λ. The following statements are equivalent:

  1. XXX is IGFR: ggg is weakly increasing on {ξ:Φ(ξ)<1}\{\xi : \Phi(\xi) < 1\}{ξ:Φ(ξ)<1}.
  2. XLX_LXL​ is IFR: its failure rate hLh_LhL​ is weakly increasing on {ξ:ΦL(ξ)<1}\{\xi : \Phi_L(\xi) < 1\}{ξ:ΦL​(ξ)<1}.
  3. X⪯hrλXX \preceq_{hr} \lambda XX⪯hr​λX for every λ≥1\lambda \ge 1λ≥1: h(ξ)≥hλ(ξ)h(\xi) \ge h_\lambda(\xi)h(ξ)≥hλ​(ξ) for all ξ≥0\xi \ge 0ξ≥0 with Φ(ξ)<1\Phi(\xi) < 1Φ(ξ)<1.
  4. f(ξ,θ)=Φˉ(ξ/θ)f(\xi,\theta) = \bar\Phi(\xi/\theta)f(ξ,θ)=Φˉ(ξ/θ) is TP₂ on (0,∞)2(0,\infty)^2(0,∞)2:
Φˉ(ξ1/θ1) Φˉ(ξ2/θ2)−Φˉ(ξ1/θ2) Φˉ(ξ2/θ1)≥0for 0<ξ1<ξ2, 0<θ1<θ2.\bar\Phi(\xi_1/\theta_1)\,\bar\Phi(\xi_2/\theta_2) - \bar\Phi(\xi_1/\theta_2)\,\bar\Phi(\xi_2/\theta_1) \ge 0 \quad\text{for } 0<\xi_1<\xi_2,\ 0<\theta_1<\theta_2.Φˉ(ξ1​/θ1​)Φˉ(ξ2​/θ2​)−Φˉ(ξ1​/θ2​)Φˉ(ξ2​/θ1​)≥0for 0<ξ1​<ξ2​, 0<θ1​<θ2​.

The IGFR condition makes the revenue and newsvendor-contract problems of operations management unimodal, but it is hard to check directly. Theorem 1 replaces it by an IFR condition on log⁡X\log XlogX, by a hazard-rate comparison under scaling, or by a total-positivity condition, each of which brings the known theory of IFR distributions to bear.

Formalization Note XXX is represented by its law μ\muμ on R\mathbb RR with μ((−∞,0))=0\mu((-\infty,0)) = 0μ((−∞,0))=0; XLX_LXL​ and λX\lambda XλX are the image laws under the measurable maps log⁡\loglog and x↦λxx \mapsto \lambda xx↦λx. The density is pinned to its regular version, because IFR and IGFR read it pointwise and the theorem is false for an arbitrary version. The rates are read on {Φ<1}\{\Phi < 1\}{Φ<1} (where Φˉ>0\bar\Phi > 0Φˉ>0), and the hazard-rate order on ξ≥0\xi \ge 0ξ≥0 with Φ(ξ)<1\Phi(\xi) < 1Φ(ξ)<1, where the paper's hhh is finite. Part 4 is stated on (0,∞)2(0,\infty)^2(0,∞)2: θ>0\theta > 0θ>0 is forced by ξ/θ\xi/\thetaξ/θ, and ξ≤0\xi \le 0ξ≤0 adds nothing since Φˉ(ξ/θ)=1\bar\Phi(\xi/\theta) = 1Φˉ(ξ/θ)=1 there.

Preamble
import Mathlib
import Definitions.Def_LariviereIGFR_Char_Setting
Formal statement
namespace LariviereIGFR.Char

open MeasureTheory ProbabilityTheory

theorem theorem_1 (μ : Measure ℝ) [IsProbabilityMeasure μ] (φ : ℝ → ℝ)
    (hnn : μ (Set.Iio 0) = 0) (hφ : IsRegDensity μ φ) :
    List.TFAE
      [ IsIGFR μ φ,
        IsIFR (μ.map Real.log) (fun ξ => Real.exp ξ * φ (Real.exp ξ)),
        ∀ c : ℝ, 1 ≤ c → HazardRateLE μ φ (μ.map (fun x => c * x)) (fun ξ => φ (ξ / c) / c),
        IsTP2On (fun ξ θ => survival μ (ξ / θ)) (Set.Ioi 0) (Set.Ioi 0) ] := by sorry

end LariviereIGFR.Char
Source
Lariviere, A note on probability distributions with increasing generalized failure rates, Oper. Res. 54(3) (2006), p. 603, Theorem 1
Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 7, 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