Footnote 18 and Lemma 5.4 — Hellinger divergence is at most KL
ProvedDROReview.PhiDiv.hellinger_le_kldistributionally-robust-optimizationhellingerkullback-leiblerp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1phi-divergence
For any probability measures and on the same measurable space, the Hellinger divergence does not exceed the Kullback–Leibler divergence:
This is the second comparison in (5.10) and is repeated explicitly in footnote 18. It implies that every KL ambiguity ball lies in the Hellinger ball of the same radius.
Formalization Note Both divergences take values in . No absolute-continuity assumption is imposed; the extended divergence handles singular pairs.
Preamble
import Definitions.Def_DROReview_PhiDiv_phiDiv import Definitions.Def_DROReview_PhiDiv_generators
Formal statement
namespace DROReview.PhiDiv
open MeasureTheory
/-- Footnote 18, p. 35, and Lemma 5.4, (5.10), second inequality, p. 34. -/
theorem hellinger_le_kl {Ξ : Type*} [MeasurableSpace Ξ]
(P P₀ : ProbabilityMeasure Ξ) :
phiDiv phiH P P₀ ≤
phiDiv phiKL P P₀ := by sorry
end DROReview.PhiDiv
Source
Rahimian, Mehrotra, Distributionally Robust Optimization: A Review, arXiv:1908.05659v1 (2019), p. 34, Lemma 5.4, (5.10), second inequality; p. 35, footnote 18
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.