Logarithmic derivative as a zero sum and correction
ProvedZeroFreeBridge.logDeriv_eq_herglotz_add_entirearda-rh65k-transfernumber-theoryriemann-zeta
Let , and suppose and have the defined canonical decomposition on , are analytic near the disk, and is zero-free there. Let be the finite support of in the disk, and take in the disk outside . With ,
The first sum records the zeros; the second is the canonical-factor correction.
Preamble
import Mathlib
import Theorems.Thm_ZeroFreeBridge_logDeriv_canonicalFactor
import Theorems.Thm_ZeroFreeBridge_logDeriv_split_off_zeros
/- PHASE 4 (dVP frontier, BLASCHKE item (d4) core): the ζ log-derivative in Herglotz + entire-part
form, with the canonical (Blaschke) factors.
Expanding `DlvpBlaschkeSplit.logDeriv_split_off_zeros` via `DlvpCanonicalLogDeriv.logDeriv_
canonicalFactor` and `n u = -divisor u` gives the dVP explicit-formula shape:
`logDeriv ζ z = Σ_ρ (divisor ρ)/(z-ρ) + [ Σ_ρ (divisor ρ)·conj ρ/(R²-conj ρ·z) + logDeriv g z ]`,
i.e. `ζ'/ζ = Z + E` with `Z` the Herglotz zero-sum (Re ≥ 0 near the line, droppable) and the entire
part `E` = correction + `logDeriv g` bounded `O(L)` by `DlvpCorrectionBound` + item (c). This is
exactly the `w = Z + E` input to the reduction skeleton `DlvpBCSum`. conjecture1_proved = False.
-/
open Complex Metric MeromorphicOn
section
namespace ZeroFreeBridge
end ZeroFreeBridge
open ZeroFreeBridge
Formal statement
theorem ZeroFreeBridge.logDeriv_eq_herglotz_add_entire {f g : ℂ → ℂ} {R : ℝ} (hR : 0 < R)
(D : CanonicalDecomp f g R)
(hf_ana : AnalyticOnNhd ℂ f (ball 0 R)) (hg_ana : AnalyticOnNhd ℂ g (ball 0 R))
(hg_ne : ∀ w ∈ ball (0 : ℂ) R, g w ≠ 0)
(hfin : (Function.support (fun u => -(divisor f (ball 0 R) u))).Finite)
{z : ℂ} (hz : z ∈ ball (0 : ℂ) R) (hzne : ∀ u ∈ hfin.toFinset, z ≠ u) :
logDeriv f z
= (∑ u ∈ hfin.toFinset, (divisor f (ball 0 R) u : ℂ) / (z - u))
+ ((∑ u ∈ hfin.toFinset,
(divisor f (ball 0 R) u : ℂ) * (starRingEnd ℂ) u
/ ((R : ℂ) ^ 2 - (starRingEnd ℂ) u * z))
+ logDeriv g z) := by sorry
Source