AFTD Auto-Formalizing Theoretical Domains

redistribution_analytic_disk_log_submean

theoremverified

For a complex function analytic on a neighbourhood of a closed disk of positive radius, with nonzero value at its centre, the logarithm of the central norm is at most the circle average of the logarithmic norm.

Statement

theorem redistribution_analytic_disk_log_submean
    (f : ℂ → ℂ) (c : ℂ) (R : ℝ) (hR : 0 < R)
    (hf : AnalyticOnNhd ℂ f (closedBall c R)) (hfc : f c ≠ 0) :
    Real.log ‖f c‖ ≤ Real.circleAverage (fun z => Real.log ‖f z‖) c R
Source
Uniform helper for IN-251, the proof of arXiv:2610.10995 Lemmas4.2/4.3. The disk logarithmic mean theorem and logarithm series already exist in the installed Mathlib; the upper-half-plane boundary passage, singular cosine integral and main harmonic bound remain unresolved. No research advance.
Verified
27 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Used by 1

The proof, explained

The proof applies Jensen's formula relating the circle average of the logarithmic norm of an analytic function f to log ‖f(c)‖ plus a sum of log(R / ‖c - u‖) weighted by the divisor (zero multiplicities) of f on the disk. Because f is analytic on the closed disk, all multiplicities are non-negative and poles do not occur. For zeros inside the disk, the distance to the centre is at most R, so the logarithmic factors are also non-negative, while all other terms vanish. Hence the sum is non-negative, yielding the submean inequality.

  1. f is analytic on the closed disk of radius |R| centered at c.

    Since the radius R is strictly positive (hR), |R| = R, so analyticity on the closed ball of radius R directly implies analyticity on the closed ball of radius |R|.

    lines 7–8

    uses abs_of_pos

      have hfa : AnalyticOnNhd ℂ f (closedBall c |R|) := by
        simpa only [abs_of_pos hR] using hf
  2. The inequality reduces to showing that the divisor sum is non-negative: 0 ≤ ∑ᶠ u, divisor f (closedBall c |R|) u * log (R * ‖c - u‖⁻¹).

    By Jensen's formula (the circle average formula for log norms of analytic functions), the circle average of log ‖f z‖ equals log ‖f c‖ plus the weighted sum over all points u of the divisor of f times log (R / ‖c - u‖). Linear arithmetic then reduces the submean inequality log ‖f c‖ ≤ circleAverage ... to showing that this divisor sum is non-negative.

    lines 9–11

    uses ne_of_gt

      rw [hfa.circleAverage_log_norm (ne_of_gt hR) hfc]
      suffices 0 ≤ ∑ᶠ u, divisor f (closedBall c |R|) u *
          Real.log (R * ‖c-u‖⁻¹) by linarith
  3. It suffices to show that each summand for an arbitrary point u is non-negative.

    A finitely supported sum of real numbers is non-negative whenever every individual term is non-negative, as provided by finsum_nonneg. We fix an arbitrary point u.

    lines 12–13

    uses finsum_nonneg

      apply finsum_nonneg
      intro u
  4. For u in the closed ball with u = c, the summand is non-negative (indeed zero).

    We split into cases on whether u is in the closed ball. When u is in the closed ball and u = c, the divisor of f at c equals the analytic order of vanishing of f at c. Since f(c) ≠ 0, this order is 0, so the summand evaluates to 0 * log(...) = 0 ≥ 0.

    lines 14–19

      by_cases hu : u ∈ closedBall c |R|
      · by_cases he : u=c
        · subst u
          rw [hfa.divisor_apply (by simp),
            (hfa c (by simp)).analyticOrderAt_eq_zero.mpr hfc]
          simp
  5. For u in the closed ball with u ≠ c, each factor of the summand is non-negative.

    Because f is analytic on the disk, its divisor (the order of zeros) is non-negative at every point. For u ≠ c, ‖c - u‖ > 0. Since u is in the closed ball of radius R, ‖c - u‖ ≤ R, so the quotient R / ‖c - u‖ is at least 1. Taking logarithms preserves this inequality, so log(R * ‖c - u‖⁻¹) ≥ 0. The product of two non-negative terms is non-negative.

    lines 20–26

    uses mul_nonneg, Real.log_nonneg, norm_pos_iff.mpr, sub_ne_zero.mpr, Ne.symm, mem_closedBall, dist_eq_norm', abs_of_pos, div_eq_mul_inv, one_le_div

        · apply mul_nonneg (by exact_mod_cast hfa.divisor_nonneg u)
          apply Real.log_nonneg
          have hnu : 0 < ‖c-u‖ := norm_pos_iff.mpr (sub_ne_zero.mpr (Ne.symm he))
          have hle : ‖c-u‖ ≤ R := by
            simpa only [mem_closedBall, dist_eq_norm', abs_of_pos hR] using hu
          rw [← div_eq_mul_inv]
          exact (one_le_div hnu).mpr hle
  6. For u outside the closed ball, the summand is zero.

    By definition, the divisor of f restricted to the closed disk vanishes outside the disk. Hence the divisor evaluates to 0 and the summand is 0 ≥ 0.

    lines 27–28

      · rw [(divisor f (closedBall c |R|)).apply_eq_zero_of_notMem hu]
        simp
abs_of_pos
For any positive real number a, the absolute value |a| equals a.
ne_of_gt
If a > b, then a ≠ b.
finsum_nonneg
If every term in a family is non-negative, their finitely supported sum (finsum) is non-negative.
mul_nonneg
The product of two non-negative numbers is non-negative.
Real.log_nonneg
If x ≥ 1, then Real.log x ≥ 0.
norm_pos_iff.mpr
If an element is non-zero, its norm is strictly positive.
sub_ne_zero.mpr
If a ≠ b, then a - b ≠ 0.
Ne.symm
If a ≠ b, then b ≠ a.
mem_closedBall
A point x is in the closed ball of centre c and radius r if and only if dist x c ≤ r.
dist_eq_norm'
The distance between a and b in a normed space is ‖b - a‖.
div_eq_mul_inv
Division a / b is definitionally multiplication by the reciprocal, a * b⁻¹.
one_le_div
For b > 0, the quotient a / b is at least 1 if and only if b ≤ a.

Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

open Finset Real MeasureTheory Metric MeromorphicOntheorem redistribution_analytic_disk_log_submean    (f : ℂ → ℂ) (c : ℂ) (R : ℝ) (hR : 0 < R)    (hf : AnalyticOnNhd ℂ f (closedBall c R)) (hfc : f c ≠ 0) :    Real.log ‖f c‖ ≤ Real.circleAverage (fun z => Real.log ‖f z‖) c R := by  have hfa : AnalyticOnNhd ℂ f (closedBall c |R|) := by    simpa only [abs_of_pos hR] using hf  rw [hfa.circleAverage_log_norm (ne_of_gt hR) hfc]  suffices 0 ≤ ∑ᶠ u, divisor f (closedBall c |R|) u *      Real.log (R * ‖c-u‖⁻¹) by linarith  apply finsum_nonneg  intro u  by_cases hu : u ∈ closedBall c |R|  · by_cases he : u=c    · subst u      rw [hfa.divisor_apply (by simp),        (hfa c (by simp)).analyticOrderAt_eq_zero.mpr hfc]      simp    · apply mul_nonneg (by exact_mod_cast hfa.divisor_nonneg u)      apply Real.log_nonneg      have hnu : 0 < ‖c-u‖ := norm_pos_iff.mpr (sub_ne_zero.mpr (Ne.symm he))      have hle : ‖c-u‖ ≤ R := by        simpa only [mem_closedBall, dist_eq_norm', abs_of_pos hR] using hu      rw [← div_eq_mul_inv]      exact (one_le_div hnu).mpr hle  · rw [(divisor f (closedBall c |R|)).apply_eq_zero_of_notMem hu]    simp
See it in the graphDiscuss this resultChallenge it