Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

(2.47), p. 767 — for h = g′(x/n), sup_{u≥0} e^{hu−g(u)} = e^{hx/n−g(x/n)}

Proved
NagaevLD.GenMoment.eq_2_47

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

convexitylarge-deviationsp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Let g:R→Rg:\mathbb R\to\mathbb Rg:R→R have, on [0,∞)[0,\infty)[0,∞), a positive nondecreasing derivative g′g'g′. Let n≥1n\ge 1n≥1 be an integer, x>0x>0x>0, and h=g′(x/n)h=g'(x/n)h=g′(x/n). Then the function u↦ehu−g(u)u\mapsto e^{hu-g(u)}u↦ehu−g(u) on u≥0u\ge 0u≥0 attains its maximum at u=x/nu=x/nu=x/n:

sup⁡u≥0ehu−g(u)=ehx/n−g(x/n),\sup_{u\ge 0}e^{hu-g(u)}=e^{hx/n-g(x/n)},u≥0sup​ehu−g(u)=ehx/n−g(x/n),

and the supremum is a maximum.

With f(u)=hu−g(u)f(u)=hu-g(u)f(u)=hu−g(u) one has f′(u)=h−g′(u)f'(u)=h-g'(u)f′(u)=h−g′(u), which vanishes at u=x/nu=x/nu=x/n; this identifies the supremum appearing in (2.46).

Formalization Note The supremum is stated as IsGreatest of the image of [0,∞)[0,\infty)[0,∞), which says both that the value is attained at x/nx/nx/n and that it bounds every other value; a real sSup is avoided because it is 000 on unbounded sets. The hypotheses on ggg are imposed on [0,∞)[0,\infty)[0,∞) only (including a two-sided derivative at 000), the only region where the paper evaluates ggg; the paper's hypothesis that ggg has a positive nondecreasing derivative implies them. The implicit n≥1n\ge1n≥1 of the paper is explicit in Lean.

Preamble
import Mathlib
import Definitions.Def_NagaevLD_GenMoment_Setting

open MeasureTheory ProbabilityTheory
Formal statement
namespace NagaevLD.GenMoment

/-- (2.47), p. 767: for `h = g′(x/n)` the supremum of `e^{hu - g(u)}` over `u ≥ 0` is attained
at `u = x/n` and equals `e^{hx/n - g(x/n)}`. -/
theorem eq_2_47 (g g' : ℝ → ℝ) (hg : ∀ u : ℝ, 0 ≤ u → HasDerivAt g (g' u) u)
    (hg'pos : ∀ u : ℝ, 0 ≤ u → 0 < g' u) (hg'mono : MonotoneOn g' (Set.Ici 0))
    (n : ℕ) (hn : 0 < n) (x : ℝ) (hx : 0 < x) (h : ℝ) (hh : h = g' (x / n)) :
    IsGreatest ((fun u : ℝ => Real.exp (h * u - g u)) '' Set.Ici 0)
      (Real.exp (h * (x / n) - g (x / n))) := by sorry

end NagaevLD.GenMoment
Source
Nagaev, Large deviations of sums of independent random variables, Ann. Probab. 7 (1979), p. 767, proof of Theorem 2.5, (2.47)
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me