p. 751 — for B > 0, A > 0: condition (1.4) ⇔ h₁ ≤ max[t/y, h₂], h₁ = αx/e^tB, h₂ = (1/y)log(βxy^{t−1}/A + 1)
ProvedNagaevLD.FukNagaev.cond_1_4_ifffuk-nagaev-inequalitylarge-deviationsp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let , , , , , and let , stand for and . Put
Then condition (1.4), , holds if and only if
and are the minimizers of the two convex parts and of the exponent in (1.18). The equivalence translates the case analysis of the proof into the conditions (1.4) and (1.6) of the theorem.
Formalization Note The page writes "the condition is equivalent to (1.4)", right after proving (1.19) "for "; (1.4) is a non-strict inequality, so the strict sign is a misprint and the statement uses . It is stated for arbitrary positive scalars , where are finite; the case is the separate disjunct of cond_1_4.
Preamble
import Mathlib import Definitions.Def_NagaevLD_FukNagaev_Setting
Formal statement
namespace NagaevLD.FukNagaev
open MeasureTheory ProbabilityTheory
theorem cond_1_4_iff (t : ℝ) (ht : 2 ≤ t) (α β : ℝ) (hα0 : 0 < α) (hα1 : α < 1)
(hβ : β = 1 - α) (x : ℝ) (hx : 0 < x) (y : ℝ) (hy0 : 0 < y) (B A : ℝ) (hB : 0 < B)
(hA : 0 < A) :
cond_1_4 t α β x y B A ↔
α * x / (Real.exp t * B) ≤
max (t / y) (1 / y * Real.log (β * x * y ^ (t - 1) / A + 1)) := by sorry
end NagaevLD.FukNagaev
Source
Nagaev, Large deviations of sums of independent random variables, Ann. Probab. 7 (1979), pp. 750–751, proof of Theorem 1.3, definition of h₁, h₂ and the sentence after (1.19)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.