Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 5.3 — with a random threshold G, the shock survival probabilities are submultiplicative for all F iff G is NBU; then H is NBU

Proved
ShockWear.RandThreshold.theorem53

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

life-distributionsnbup2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1reliabilityshock-models

A device receives shocks; the iiith shock causes damage XiX_iXi​, the XiX_iXi​ are independent with common distribution FFF, and the device fails as soon as the accumulated damage exceeds a random threshold YYY with distribution GGG, independent of the damages. Assume F(z)=G(z)=0F(z) = G(z) = 0F(z)=G(z)=0 for z<0z < 0z<0. The probability of surviving kkk shocks is

Pˉk=∫0∞F(k)(x) dG(x),k=0,1,…,(5.1)\bar P_k = \int_0^\infty F^{(k)}(x)\, dG(x), \qquad k = 0, 1, \dots, \tag{5.1}Pˉk​=∫0∞​F(k)(x)dG(x),k=0,1,…,(5.1)

where F(k)F^{(k)}F(k) is the kkk-fold convolution of FFF (F(0)F^{(0)}F(0) degenerate at 000). If shocks arrive as a Poisson process of rate λ>0\lambda > 0λ>0, the device's life distribution HHH has survival function

Hˉ(t)=∑k=0∞e−λt(λt)kk!∫0∞F(k)(x) dG(x),t≥0.(5.2)\bar H(t) = \sum_{k=0}^\infty e^{-\lambda t}\frac{(\lambda t)^k}{k!}\int_0^\infty F^{(k)}(x)\, dG(x), \qquad t \ge 0. \tag{5.2}Hˉ(t)=k=0∑∞​e−λtk!(λt)k​∫0∞​F(k)(x)dG(x),t≥0.(5.2)

Theorem.

  1. Pˉj+k≤PˉjPˉk\bar P_{j+k} \le \bar P_j \bar P_kPˉj+k​≤Pˉj​Pˉk​ for all j,k=0,1,…j, k = 0, 1, \dotsj,k=0,1,… and for all FFF with F(z)=0F(z) = 0F(z)=0 for z<0z < 0z<0 if and only if GGG is NBU, i.e. Gˉ(s+t)≤Gˉ(s)Gˉ(t)\bar G(s + t) \le \bar G(s)\bar G(t)Gˉ(s+t)≤Gˉ(s)Gˉ(t) for all s,t≥0s, t \ge 0s,t≥0, where Gˉ=1−G\bar G = 1 - GGˉ=1−G.
  2. If GGG is NBU, then for every such FFF and every λ>0\lambda > 0λ>0, HHH given by (5.2) is NBU: Hˉ(t+x)≤Hˉ(x)Hˉ(t)\bar H(t + x) \le \bar H(x)\bar H(t)Hˉ(t+x)≤Hˉ(x)Hˉ(t) for all x,t≥0x, t \ge 0x,t≥0.

The first part characterizes the NBU class through the cumulative-damage model with a random threshold: the threshold law is NBU exactly when submultiplicativity of the survival probabilities holds regardless of how damage is distributed. The second part is the reliability conclusion: an NBU threshold yields an NBU life distribution under Poisson shocks.

Formalization Note. Distributions are probability measures on R\mathbb RR giving mass 000 to (−∞,0)(-\infty, 0)(−∞,0); "for all FFF" ranges over every such measure, inside the equivalence. Pˉk\bar P_kPˉk​ is (5.1) exactly as printed, P{X1+⋯+Xk≤Y}P\{X_1 + \cdots + X_k \le Y\}P{X1​+⋯+Xk​≤Y}, and the integral is over R\mathbb RR, so an atom of GGG at 000 (which the theorem allows) is kept. The paper's proof uses EGˉ(X1+⋯+Xk)E\bar G(X_1 + \cdots + X_k)EGˉ(X1​+⋯+Xk​), which equals (5.1) under its convention that F(k)F^{(k)}F(k) and GGG have no common discontinuities; we do not add that convention, and the theorem holds for (5.1) without it. NBU is the cross-multiplied form of definition (v), on x,t≥0x, t \ge 0x,t≥0. λ>0\lambda > 0λ>0 is quantified in the second part.

Preamble
import Mathlib
import Definitions.Def_ShockWear_RandThreshold_Model
Formal statement
namespace ShockWear.RandThreshold

open MeasureTheory ProbabilityTheory

theorem theorem53 (ν : Measure ℝ) [IsProbabilityMeasure ν] (hν : ν (Set.Iio 0) = 0) :
    ((∀ μ : Measure ℝ, IsProbabilityMeasure μ → μ (Set.Iio 0) = 0 →
        ∀ j k : ℕ, randP μ ν (j + k) ≤ randP μ ν j * randP μ ν k) ↔ IsNBU (survOf ν)) ∧
    (IsNBU (survOf ν) → ∀ μ : Measure ℝ, IsProbabilityMeasure μ → μ (Set.Iio 0) = 0 →
        ∀ lam : ℝ, 0 < lam → IsNBU (shockSurv lam (randP μ ν))) := by sorry

end ShockWear.RandThreshold
Source
Esary, Marshall and Proschan, Shock Models and Wear Processes, Ann. Probability 1 (1973), p. 645, Theorem 5.3, with (5.1) and (5.2) on p. 642
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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