Tail-mirror inequality:
ProvedOWMR.tail_mirrorapproximation-algorithmsinventoryinventory-theoryoperations-researchrenewal-theorystochastic-control
Let be the renewal measure of i.i.d. nonnegative demands with positive mean, , and let , , , and . Let be any real number with
Then
By the occupation identity, is the long-run average mirror loss of a width- policy at depletion center , so the mirror cost of (14) satisfies the hypothesis, and the conclusion is (17).
Preamble
import Definitions.Def_owmr_renewal open MeasureTheory open scoped ENNReal NNReal
Formal statement
namespace OWMR
theorem tail_mirror (ρ : Measure ℝ≥0) [IsProbabilityMeasure ρ]
(hρ : 0 < ∫⁻ x, (x : ℝ≥0∞) ∂ρ) (h p b : ℝ) (hh : 0 < h) (hp : 0 < p) (hb : 0 < b)
(B : ℝ) (hB : ∀ z : ℝ, B * (renewalMeasure ρ (Set.Ico 0 b)).toReal
≤ ∫ x in Set.Ico 0 b, (p * max (z - x) 0 + h * max (x - z) 0) ∂(renewalMeasure ρ)) :
p * ∫ x in Set.Ico b (2 * b), (2 * b - x) ∂(renewalMeasure ρ)
+ 4 * (p / h) * (renewalMeasure ρ (Set.Ico 0 b)).toReal * B
≤ (3 * (p / h) + 1) * (h * p / (h + p)) * b * (renewalMeasure ρ (Set.Ico 0 b)).toReal := by sorry
end OWMRSource
M. Hu, "A 5/2 Guarantee for Echelon Stock (s,S) Policies in One-Warehouse Multi-Retailer Systems", working paper (manuscript of November 2026), Lemma 6 (Tail-Mirror Inequality), eq. (17), p. 18; proof on pp. 18-19.