Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Committed-supply bound quantities BμB_\muBμ​, DμD_\muDμ​ and the concave lower curve Lμ\mathcal L_\muLμ​

Definition
LostSalesRQ_lower_curve

by Ming Hu · Nov 17, 2026 · Mathlib 0df444a (Lean v4.33.1)

approximation-guaranteeinventorylost-salesoperations-research

For a lead-time demand mean μ≥0\mu\ge0μ≥0 and a rank weight η\etaη, with pi,Fi,Glp_i,F_i,G_lpi​,Fi​,Gl​ the Poisson functions of the companion definition file:

  1. The rank bound (eq. (8))
Bμ(η)=min⁡l≥0{Gl(μ)+η P(N(μ)≥l)}.B_\mu(\eta)=\min_{l\ge0}\big\{G_l(\mu)+\eta\,\mathbb P(N(\mu)\ge l)\big\}.Bμ​(η)=l≥0min​{Gl​(μ)+ηP(N(μ)≥l)}.
  1. The aging-rate supremum (eq. (9)), over the number m≥0m\ge0m≥0 of pending units and ordered residual delivery times,
Dμ(η)=sup⁡m≥0, 0≤τ1≤⋯≤τm≤μ ∑i=1m(η pi−1(τi)−Fi−1(τi)),D_\mu(\eta)=\sup_{m\ge0,\ 0\le\tau_1\le\cdots\le\tau_m\le\mu}\ \sum_{i=1}^{m}\big(\eta\,p_{i-1}(\tau_i)-F_{i-1}(\tau_i)\big),Dμ​(η)=m≥0, 0≤τ1​≤⋯≤τm​≤μsup​ i=1∑m​(ηpi−1​(τi​)−Fi−1​(τi​)),

the empty sum (m=0m=0m=0) included. 3. The concave lower curve (eq. (13))

Lμ(a)=sup⁡η≥1min⁡{Bμ(η), a−Dμ(η)}.\mathcal L_\mu(a)=\sup_{\eta\ge1}\min\{B_\mu(\eta),\,a-D_\mu(\eta)\}.Lμ​(a)=η≥1sup​min{Bμ​(η),a−Dμ​(η)}.

Proposition 1 shows that min⁡{Bμ(η),a−Dμ(η)}\min\{B_\mu(\eta),a-D_\mu(\eta)\}min{Bμ​(η),a−Dμ​(η)} bounds the zero-setup optimal cost from below for every η≥1\eta\ge1η≥1, so Lμ(a)\mathcal L_\mu(a)Lμ​(a) is the best such bound.

Formalization Note The minimum in (8) is a real infimum over l∈Nl\in\mathbb Nl∈N (bounded below by 000 for η≥0\eta\ge0η≥0). The suprema in (9) and (13) are real suprema of explicit sets that are nonempty and bounded above on the paper's domain (Dμ(η)≤η2/2−η/3D_\mu(\eta)\le\eta^2/2-\eta/3Dμ​(η)≤η2/2−η/3 and Lμ(a)≤a\mathcal L_\mu(a)\le aLμ​(a)≤a); in the Lean sum over i:Fin mi:\mathrm{Fin}\,mi:Finm the index iii carries rank i+1i+1i+1.

Definition code
import Definitions.Def_LostSalesRQ_poisson_functions

/-!
# The committed-supply lower-bound quantities

`B_μ(η)` (eq. (8)), `D_μ(η)` (eq. (9)) and the concave lower curve `𝓛_μ(a)` (eq. (13)) of
M. Hu, "A 1.932 Guarantee for (r,Q) Policies in Continuous-Review Lost-Sales Inventory Systems".
-/

namespace LostSalesRQ

open Finset

/-- Rank bound `B_μ(η) = min_{l ≥ 0} {G_l(μ) + η P(N(μ) ≥ l)}` (eq. (8)). -/
noncomputable def rankBound (μ η : ℝ) : ℝ :=
  ⨅ l : ℕ, (expectedShortfall μ l + η * poisTail μ l)

/-- Aging-rate supremum (eq. (9)): the supremum, over the number `m ≥ 0` of pending units and
ordered residual delivery times `0 ≤ τ_1 ≤ ⋯ ≤ τ_m ≤ μ`, of
`∑_{i=1}^{m} (η p_{i-1}(τ_i) - F_{i-1}(τ_i))`. Here `Fin m` indexes the ranks `1, …, m`
from `0`, so index `i` carries rank `i + 1`. -/
noncomputable def agingRateSup (μ η : ℝ) : ℝ :=
  sSup {x : ℝ | ∃ (m : ℕ) (τ : Fin m → ℝ), Monotone τ ∧ (∀ i, 0 ≤ τ i ∧ τ i ≤ μ) ∧
    x = ∑ i : Fin m, (η * poisMass (τ i) (i : ℕ) - poisCdf (τ i) (i : ℕ))}

/-- Concave lower curve `𝓛_μ(a) = sup_{η ≥ 1} min{B_μ(η), a - D_μ(η)}` (eq. (13)). -/
noncomputable def lowerCurve (μ a : ℝ) : ℝ :=
  sSup {x : ℝ | ∃ η : ℝ, 1 ≤ η ∧ x = min (rankBound μ η) (a - agingRateSup μ η)}

end LostSalesRQ
Source
M. Hu, "A 1.932 Guarantee for (r,Q) Policies in Continuous-Review Lost-Sales Inventory Systems", arXiv:2610.34576, 2026; Section 4.3.1, eqs. (8)-(9) (pp. 11-12); Section 4.3.3, eq. (13) (p. 14).

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