Lemma 5 — P[D₁ + ⋯ + D_t ≤ η] ≤ F(η)^t
ProvedAdaptiveBaseStock.Regret.demand_sum_tailinventoryp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1probability
Let be i.i.d. nonnegative continuous demands with and distribution function , and suppose that has an infinite support. Then for every and every ,
This tail bound controls how long an initial inventory position above can survive, and it is the input of Lemma 6.
Formalization Note Only the infinite-support case of the paper's lemma is stated; the bounded-demand case is not part of this mission. The sum runs over Lean indices , which are the paper's .
Preamble
import Mathlib import Definitions.Def_AdaptiveBaseStock_Regret_Model
Formal statement
namespace AdaptiveBaseStock.Regret
open MeasureTheory ProbabilityTheory
/-- Lemma 5 (infinite-support case), p. 15: `P[D_1 + ⋯ + D_t ≤ η] ≤ F(η)^t` for every `η` and
`t ≥ 1`. -/
theorem demand_sum_tail {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(D : ℕ → Ω → ℝ) (hD : IsDemandModel P D) (hinf : HasInfiniteSupport P (D 0))
(η : ℝ) (t : ℕ) (ht : 1 ≤ t) :
P.real {ω | ∑ ℓ ∈ Finset.range t, D ℓ ω ≤ η} ≤ demandCdf P (D 0) η ^ t := by sorry
end AdaptiveBaseStock.Regret
Source
Huh, Janakiraman, Muckstadt, Rusmevichientong, An Adaptive Algorithm for Finding the Optimal Base-Stock Policy in Lost Sales Inventory Systems with Censored Demand, working paper, February 8, 2007 (published version: Math. Oper. Res., 2009, DOI 10.1287/moor.1080.0367), Lemma 5 (first case), p. 15
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.