Single-location value and bounded relative value: ,
ProvedOWMR.sl_subsolutionapproximation-algorithmsinventoryinventory-theoryoperations-researchrenewal-theorystochastic-control
Let the demand law on have positive mean, let be continuous, convex, coercive ( as ) and nonincreasing on , and let . Then there is with
the single-location value from initial position over all nonanticipative policies with auxiliary randomness, and there is a measurable with such that
This is the accounting inequality (5) of Lemma 1 (with equal to up to a constant, so ), together with the identification of as the single-location value. The monotonicity on holds for every cost the source uses: and equal times the expected backlog at negative positions.
Preamble
import Definitions.Def_owmr_single_location open MeasureTheory open scoped ENNReal NNReal
Formal statement
namespace OWMR
theorem sl_subsolution (ρ : Measure ℝ≥0) [IsProbabilityMeasure ρ]
(hρ : 0 < ∫⁻ x, (x : ℝ≥0∞) ∂ρ) (f : ℝ → ℝ) (hfc : Continuous f)
(hfcv : ConvexOn ℝ Set.univ f) (hf0 : ∀ y, 0 ≤ f y)
(hfco : ∀ C : ℝ, ∃ R : ℝ, ∀ y, R ≤ |y| → C ≤ f y)
(hfa : AntitoneOn f (Set.Iic 0)) (K : ℝ) (hK : 0 ≤ K) :
∃ J : ℝ, 0 ≤ J ∧ slValue ρ f K = ENNReal.ofReal J ∧
∃ v : ℝ → ℝ, Measurable v ∧ (∀ y, -K ≤ v y ∧ v y ≤ 0) ∧
∀ y, J + max (f y - J) 0 ≤ f y + ∫ d, v (y - (d : ℝ)) ∂ρ - v y := 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 1, Section 4.1, p. 11 (equations (5) and (6)), and Appendix A (proof of Lemma 1); used in Steps 2 and 3 of the proof of Proposition 3, Section 5.3, pp. 24-25.