Proof of Theorem 8 — for any γ > 0, the excess inventory y_t − ŷ_t is dominated by the queue Z_t(m)
ProvedAIMInventory.UnknownRho.couplingConsider the AIM algorithm in the non-perishable inventory case with arbitrary step scale , upper bound , costs and initial target . Let be the implemented order-up-to level and the target level in period , and let be the decreasing-inflow queue , , with
Then for every demand realization and every period ,
This is the version of Theorem 6 that needs no relation between and a lower bound on mean demand: the inflow of the dominating queue is itself. It converts the excess cost into a queueing quantity in the proof of Theorem 8.
Formalization Note. The paper states the inequality "with probability one"; the Lean statement is the stronger pathwise form, for every demand path, which is what the paper's argument gives. Demand values are nonnegative reals, and the paper's is d (t-1).
import Mathlib import Definitions.Def_AIMInventory_UnknownRho_Model open MeasureTheory open scoped ENNReal NNReal
namespace AIMInventory.UnknownRho
/-- The pathwise coupling in the proof of Theorem 8, p. 13. -/
theorem coupling (c : Parameters) (d : DemandPath) (t : ℕ) (ht : 1 ≤ t) :
aimLevel c d t - aimTarget c d t ≤ Z d (m c) t := by sorry
end AIMInventory.UnknownRho
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.