Proof of Lemma 1, p. ec6 — every ξ ∈ Ξ(ℓ) lies in Ξ_ε(ℓ) for all small ε, so Ξ_ε(ℓ) ≠ ∅
ProvedKAdaptability.EpsApprox.lemma_1_nonemptyinteger-programmingp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1robust-optimization
Fix a decision and an index vector , and let and be the exact and the approximate uncertainty sets of problems (6) and .
For every there is such that
In particular, if , then for all sufficiently small . (The paper takes ; when no is nonzero, for every and any works.)
This is the first step of the proof of Lemma 1 and the reason the effective domains of (6) and agree for small .
Preamble
import Mathlib import Definitions.Def_KAdaptability_EpsApprox_Approx
Formal statement
namespace KAdaptability.EpsApprox
open Problem
/-- **Proof of Lemma 1, first step** (p. ec6). Fix a decision `(x, {y^k}_{k∈𝒦}) ∈ 𝒳 × 𝒴^K` and
`ℓ ∈ ℒ`. Every `ξ ∈ Ξ(ℓ)` lies in `Ξ_ε(ℓ)` for all `ε ∈ (0, ε′]`, for some `ε′ > 0`; in particular
`Ξ_ε(ℓ) ≠ ∅` for all sufficiently small `ε > 0` whenever `Ξ(ℓ) ≠ ∅`. -/
theorem lemma_1_nonempty {N M L nQ R K : ℕ} (P : Problem N M L nQ R)
(x : Fin N → ℝ) (y : Fin K → Fin M → ℝ) (hd : P.IsDecision x y)
(ℓ : Fin K → Fin (L + 1)) :
∀ ξ ∈ P.XiL x y ℓ, ∃ ε' > 0, ∀ ε ∈ Set.Ioc (0 : ℝ) ε', ξ ∈ P.XiEps ε x y ℓ := by sorry
end KAdaptability.EpsApprox
Source
Hanasusanto, Kuhn, Wiesemann, K-Adaptability in Two-Stage Robust Binary Programming, preprint, Optimization Online 2014/03/4294 (revision of 2015-03-30), p. ec6 (PDF p. 40), proof of Lemma 1, first paragraph
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.