Proof of Corollary 1, Step 2 (pp. 28–29) — (A1), (A2), (A4) and I < ∞ give (P-USC)
ProvedModelRiskOT.PrimalOpt.corollary_1_step_2optimal-transportp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1uniform-integrabilityweak-convergence
Let be a real normed space that is locally compact in its norm topology, with its Borel -algebra, let be a probability measure on and . Let and satisfy (A1), (A2) and (A4), and assume . Then has Property (P-USC): for every sequence with ,
This is the second verification in the proof of Corollary 1; it rests on a uniform integrability property of under monotone feasible plans.
Formalization Note The finiteness is the part of the standing assumption the step uses. As in Step 1, PolishSpace E is redundant for a locally compact real normed space.
Preamble
import Mathlib import Definitions.Def_ModelRiskOT_PrimalOpt_PUSC import Definitions.Def_ModelRiskOT_PrimalOpt_GrowthAssumptions
Formal statement
namespace ModelRiskOT.PrimalOpt
open MeasureTheory
/-- **Proof of Corollary 1, Step 2** (pp. 28–29): on a real normed space `E` that is locally
compact, (A1), (A2), (A4), `δ > 0` and `I < ∞` imply (P-USC). -/
theorem corollary_1_step_2 {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E]
[LocallyCompactSpace E] [PolishSpace E] [MeasurableSpace E] [BorelSpace E]
(c : E → E → ℝ) (f : E → ℝ) (μ : Measure E) [IsProbabilityMeasure μ]
(δ : ℝ) (hδ : 0 < δ) (hA1 : ModelRiskOT.Duality.AssumptionA1 c) (hA2 : AssumptionA2 f μ)
(hA4 : AssumptionA4 c f) (hfin : ModelRiskOT.Duality.primalValue c f μ δ < ⊤) :
PUSC c f μ δ := by sorry
end ModelRiskOT.PrimalOpt
Source
Blanchet & Murthy, Quantifying Distributional Model Risk via Optimal Transport, arXiv:1604.01446v2, pp. 28–29, proof of Corollary 1, Step 2, Eq. (40)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.