Proposition 2: exact penalty calibration
ProvedLostSalesRQ.penalty_calibrationapproximation-guaranteeinventorylost-salesoperations-research
Fix . For let , where is a rate-one Poisson process. Then
and is a continuous, strictly increasing bijection from onto itself.
At the penalty the two terms of the committed-supply bound coincide, so every penalty has a unique calibrated weight at which Proposition 1 gives exactly .
Preamble
import Definitions.Def_LostSalesRQ_lower_curve import Definitions.Def_LostSalesRQ_poisson_process
Formal statement
namespace LostSalesRQ
theorem penalty_calibration (μ : ℝ) (hμ : 0 ≤ μ) :
(∀ η : ℝ, 1 ≤ η → rankBound μ η + agingRateSup μ η = calibratedPenalty μ η) ∧
ContinuousOn (calibratedPenalty μ) (Set.Ici 1) ∧
StrictMonoOn (calibratedPenalty μ) (Set.Ici 1) ∧
Set.BijOn (calibratedPenalty μ) (Set.Ici 1) (Set.Ici 1) := by sorry
end LostSalesRQSource
M. Hu, "A 1.932 Guarantee for (r,Q) Policies in Continuous-Review Lost-Sales Inventory Systems", arXiv:2610.34576, 2026; Proposition 2, eqs. (11)-(12) (p. 13); proof in Appendix C (pp. A5-A7).
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.