Exhaustion of a finite divisor by a zero list
ProvedBoxLocalization.exhaustion_by_countarda-rh65k-transfernumber-theoryriemann-zeta
Let be finite sets of complex numbers, , and let satisfy for all . If
then and for every . Thus distinct points exhaust a total multiplicity of .
Preamble
import Definitions.Def_ArdaRH65K_XiLineZeros
import Mathlib
/- Task 6 (Stage 3): RH-IN-A-BOX LOCALIZATION CAPSTONE.
Box `B = [2/5, 3/5] x [10, 35]`. This file composes the three kernel results of the project into
the crowning statement: EVERY nontrivial zero of `riemannZeta` in `B` lies on the critical line
`Re = 1/2` (and is simple). This is a Turing-style VERIFICATION OF RH INSIDE THIS BOX. It is NOT
a proof of the Riemann Hypothesis. `conjecture1_proved = False`.
The pieces composed:
1. TOTAL COUNT = 5 (kernel argument principle):
`BoxArgPrincipleZeta.box_arg_principle_zeta'` concludes `∃ s d, Σ_{ρ∈s} (d ρ) = N`, where `s`
is the ACTUAL zeta divisor support on the ball `U = ball cB 13 ⊇ B` and `d` its multiplicity
(`MeromorphicOn.divisor riemannZeta U`), GIVEN the honest Arb bundle `hArb` at `N = 5`. The
winding integer `N = 5`, edge non-vanishing, strict interiority, and the value enclosures are
the documented NON-KERNEL Arb inputs (same trust boundary as Stage 1's enclosures).
2. ON-LINE LOWER BOUND >= 5 (kernel, Stage 1):
`XiLineZeros.lambda_five_zeros_10_35` yields 5 DISTINCT `completedRiemannZeta` zeros on the line
`Re = 1/2` with imaginary parts in `(10, 35)` (from Arb-certified sign changes).
3. LAMBDA <-> ZETA BRIDGE (kernel, this file):
In `B` (`Re > 0`, so `s ≠ 0`), `riemannZeta s = completedRiemannZeta s / Gammaℝ s` with
`Gammaℝ s ≠ 0`; hence `riemannZeta s = 0 ↔ completedRiemannZeta s = 0`. The 5 on-line
`completedRiemannZeta`-zeros ARE 5 distinct on-line `riemannZeta`-zeros in `B`.
COUNTING ARGUMENT (kernel, this file): the divisor is nonnegative (zeta is analytic on `U`, no
poles), so `d ρ ≥ 1` at every zero. The 5 distinct on-line zeros lie in the support `s`, each
with `d ≥ 1`, summing to `≥ 5`. Since the TOTAL is exactly `5`, they EXHAUST the divisor: the
support equals those 5 points and each multiplicity is exactly 1. Therefore every zeta-zero in
`B` is one of the 5 on-line zeros -- so it lies on `Re = 1/2` and is simple.
conjecture1_proved = False. -/
open Complex Real
section
namespace BoxLocalization
end BoxLocalization
open BoxLocalization
/-! ## Part A: the pure combinatorial localization lemma (kernel).
If a finite multiset-with-multiplicity `d ≥ 1` on a support `s` sums to `n`, and we already exhibit
`n` DISTINCT elements of `s`, then `s` is EXACTLY those `n` elements and every multiplicity is `1`.
This is the load-bearing counting step; it is fully general and kernel-proven. -/
Formal statement
theorem BoxLocalization.exhaustion_by_count {s T : Finset ℂ} {d : ℂ → ℤ} {n : ℕ}
(hTsub : T ⊆ s) (hTcard : T.card = n)
(hd1 : ∀ ρ ∈ s, (1 : ℤ) ≤ d ρ)
(hsum : (∑ ρ ∈ s, d ρ) = (n : ℤ)) :
s = T ∧ ∀ ρ ∈ s, d ρ = 1 := by sorry
Source