Zeta logarithmic-derivative decomposition in a ball
ProvedRHInBoxAnalytic.zeta_blaschke_split_ballarda-rh65k-transfernumber-theoryriemann-zeta
Let , and suppose a closed rectangle lies inside while lies outside that open ball. There is a function differentiable on the rectangle such that the ball's finite zeta-zero support contains every zeta zero in the ball, each listed divisor multiplicity is at least one, and at every nonzero point of zeta in the ball,
The sum is over the defined finite zero support in .
Preamble
import Definitions.Def_ArdaRH65K_Turing121_RHInBoxAnalytic
import Mathlib
/- Task 2 (parameterized RH-in-a-box): generalize the zeta Blaschke split + box argument
principle from the merged FIXED ball `ball cB 13` / fixed corners to an ARBITRARY box
`[sigma0, sigma1] x [T0, T1]` with a VARIABLE Blaschke ball `(c, R)`, the geometry entering
only as hypotheses `hRpos`, `hbox_ball`, `hs1` (and the box-ordering `hsig`, `hT`).
This file reproduces, over variable corners and a variable ball, the three merged atoms
- the per-pole interior winding `Bd((z-ρ)^{-1}) = 2*pi*I` (branch-split primitive),
- the residue-sum linearity `Bd(∑ d/(z-ρ)) = 2*pi*I*∑ d`,
- the analytic E-arg-principle `Bd(E) = 0` (Mathlib `integral_boundary_rect_eq_zero...`),
and the KERNEL-DERIVED local Blaschke split of `logDeriv riemannZeta` over the variable ball
(`MeromorphicOn.extract_zeros_poles`, E-holomorphicity in-kernel), then composes them into
`RHInBoxAnalytic.zeta_count_eq_winding_generic`:
the actual divisor of `riemannZeta` over `ball c R` is nonneg, captures every box zero, and its
total equals the Arb winding integer `N` (derived from `hwind` via the argument principle).
Divisor-support finiteness is obtained from `IsCompact.inter_riemannZetaZeros_finite`
(compact `closedBall c R` meets `riemannZetaZeros` finitely), NOT from meromorphy on the closed
ball -- so no assumption is needed about the pole `1` on the closed-ball boundary; only
`1 ∉ ball c R` (open) enters, exactly as the brief requires.
conjecture1_proved = False. This is a kernel derivation of a parameterized argument-principle
count; it does NOT prove RH. -/
open Complex Filter MeasureTheory Real
open scoped Topology
section
namespace RHInBoxAnalytic
end RHInBoxAnalytic
open RHInBoxAnalytic
/-! ## Part A: function-agnostic log-derivative transfer across a codiscrete equality.
(Verbatim from the merged `BlaschkeBox.lean`; pure Mathlib, function-agnostic.) -/
/-! ## Part B: log-derivative of the factorized-rational product. (Verbatim.) -/
/-! ## Part C: `riemannZeta` meromorphic-order facts on `{1}ᶜ`. (Verbatim from merged.) -/
/-! ## Part D: generic per-pole interior winding and residue-sum atoms over variable corners.
Generalizes `WindingCount.winding_lambda_five_rect_winding` and
`BoxArgPrinciple.box_residue_sum_lambda` from the fixed corners `2/5,3/5,10,35` to variable
`sigma0 < ρ.re < sigma1`, `T0 < ρ.im < T1`. -/
/-! ## Part E: variable-ball local Blaschke split of `logDeriv riemannZeta`.
Generalizes `BlaschkeBox.zeta_blaschke_split_box` to an arbitrary ball `(c, R)` whose interior
avoids `1` (`hs1`) and whose box `[σ0,σ1]x[T0,T1] ⊆ ball c R` (`hbox_ball`). Divisor
finiteness comes from `IsCompact.inter_riemannZetaZeros_finite`. -/
Formal statement
theorem RHInBoxAnalytic.zeta_blaschke_split_ball
(sigma0 sigma1 T0 T1 : ℝ) (c : ℂ) (R : ℝ)
(hRpos : 0 < R)
(hbox_ball : ∀ ρ : ℂ, (sigma0 ≤ ρ.re ∧ ρ.re ≤ sigma1) → (T0 ≤ ρ.im ∧ ρ.im ≤ T1) →
ρ ∈ Metric.ball c R)
(hs1 : (1 : ℂ) ∉ Metric.ball c R) :
∃ (E : ℂ → ℂ),
DifferentiableOn ℂ E (Set.Icc sigma0 sigma1 ×ℂ Set.Icc T0 T1) ∧
(∀ ρ ∈ zeroFinset c R hs1,
(1 : ℤ) ≤ (MeromorphicOn.divisor riemannZeta (Metric.ball c R) : ℂ → ℤ) ρ) ∧
(∀ ρ ∈ Metric.ball c R, riemannZeta ρ = 0 → ρ ∈ zeroFinset c R hs1) ∧
(∀ z ∈ Metric.ball c R, riemannZeta z ≠ 0 →
logDeriv riemannZeta z = (∑ ρ ∈ zeroFinset c R hs1,
((MeromorphicOn.divisor riemannZeta (Metric.ball c R) : ℂ → ℤ) ρ : ℂ) / (z - ρ)) + E z) := by sorry
Source