Lemma 7.1 — good-λ principle for Φ inequalities
ProvedBurkholderDFI.ConvexPhi.lemma_7_1good-lambdamartingalesp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be nonnegative measurable random variables. Suppose , and, for every ,
If has moderate growth, and for all , and , then
This turns distribution-function estimates into expectation estimates.
Formalization Note The variables and expectations may be infinite. The case is included.
Preamble
import Mathlib import Definitions.Def_BurkholderDFI_SquareFnLp_Martingale
Formal statement
namespace BurkholderDFI.ConvexPhi
open MeasureTheory ProbabilityTheory
open scoped ENNReal NNReal
/-- Lemma 7.1, p. 26: the good-λ lemma. -/
theorem lemma_7_1 {Ω : Type*} [MeasurableSpace Ω] (P : Measure Ω) [IsProbabilityMeasure P]
(Φ : ℝ≥0∞ → ℝ≥0∞) (c : ℝ≥0) (hΦ : BurkholderDFI.SquareFnLp.IsPhi Φ c)
(f g : Ω → ℝ≥0∞) (hf : Measurable f) (hg : Measurable g)
(β δ ε : ℝ) (hβ : 1 < β) (hδ : 0 < δ) (hε : 0 < ε)
(h71 : ∀ l : ℝ, 0 < l →
P {ω | ENNReal.ofReal (β * l) < g ω ∧ f ω ≤ ENNReal.ofReal (δ * l)}
≤ ENNReal.ofReal ε * P {ω | ENNReal.ofReal l < g ω})
(γ η : ℝ≥0)
(h72 : ∀ l : ℝ, 0 < l →
Φ (ENNReal.ofReal (β * l)) ≤ γ * Φ (ENNReal.ofReal l) ∧
Φ (ENNReal.ofReal (δ⁻¹ * l)) ≤ η * Φ (ENNReal.ofReal l))
(hγε : (γ : ℝ) * ε < 1) :
∫⁻ ω, Φ (g ω) ∂P ≤ ENNReal.ofReal (γ * η / (1 - γ * ε)) * ∫⁻ ω, Φ (f ω) ∂P := by sorry
end BurkholderDFI.ConvexPhi
Source
Burkholder, Distribution Function Inequalities for Martingales, Ann. Probability 1 (1973), Lemma 7.1, p. 26
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.