Matching upper bound (sub-conjecture D1-lite, closed).
ProvedHyperbolicBerggrenGeodesics.ball_card_upperaether-cataloggeometry
Matching upper bound (sub-conjecture D1-lite, closed).
Any family of Berggren nodes inside the hyperbolic ball of radius R has at most
4 e^{2R} members. Together with hyperbolic_ball_quadratic_growth this pins the volume
growth of the Berggren tree in the hyperbolic metric to the exact exponent 2R:
e^{2R}/300 ≤ #B(R) ≤ 4 e^{2R}.
theorem HyperbolicBerggrenGeodesics.ball_card_upper{R : ℝ} (hR : 0 ≤ R) (S : Finset (ℕ × ℕ))
(hseed : ∀ p ∈ S, IsSeed p.1 p.2)
(hball : ∀ p ∈ S, ∀ h : 0 < p.1, dist (hpoint p.1 p.2 h) UpperHalfPlane.I ≤ R) :
(S.card : ℝ) ≤ 4 * Real.exp (2 * R) := by sorry
Formalization Note Transplanted verbatim from the Aether Catalog source Geometry/HyperbolicBerggrenDensity.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.
Preamble
-- Thm stub generated from Geometry/HyperbolicBerggrenDensity.lean
import Mathlib
import Definitions.Def_Geometry_HyperbolicBerggrenDensity
import Definitions.Def_Geometry_HyperbolicBerggrenGeodesics
import Definitions.Def_Geometry_HyperbolicBerggrenGeodesicsII
/-!
# Hyperbolic–Pythagorean Geodesics, cycle III: quadratic ball growth
The first cycle proved that the hyperbolic ball of radius `R` around `i` contains at least
`e^{R-2} - 1` Berggren nodes, and conjectured (sub-conjecture **C1-lite**) the true order
`e^{2R}`: the number of nodes should grow like the *hypotenuse*, not like its square root.
This file proves that conjecture. The obstruction is arithmetic, not geometric: one has to
produce quadratically many *coprime* pairs of opposite parity, which requires a sieve.
## Main results
* `card_multiples_Ioc` : the exact count of multiples of `k` in an interval `(a, b]`.
* `sum_inv_sq_odd`, `sum_inv_odd` : two telescoping estimates,
`∑_{i<n} 1/(2i+3)² ≤ 1/4` and `∑_{i<n} 1/(2i+3) ≤ √(2n+1) - 1`.
* `card_seedBox_lower` : **the sieve bound.** For `K ≥ 256` the box
`{m even, 2K < m ≤ 4K} × {n odd, 1 ≤ n ≤ 2K}` contains at least `K²/4` Euclid seeds.
* `hyperbolic_ball_quadratic_growth` : **C1-lite, closed.** For every `K ≥ 256` the
hyperbolic ball of radius `R = log K + 2` around the base point contains at least
`e^{2R}/300` distinct Berggren nodes. Since every node with hypotenuse `c` sits at
distance `≈ ½ log c`, this is the true order of growth, and it shows definitively that
geodesic search through the Berggren tree cannot beat exhaustive search: the ball that
is guaranteed to contain a colliding pair for `N` already contains `≍ N` nodes.
-/
open HyperbolicBerggrenGeodesics
open Real UpperHalfPlane
noncomputable section
/-! ## Part A. Counting multiples -/
/-! ## Part B. Two telescoping estimates -/
/-! ## Part C. The sieve -/
/-! ## Part D. From the sieve to a quadratic lower bound -/
/-! ## Part E. Quadratic volume growth of hyperbolic balls (C1-lite, closed) -/
/-! ## Part F. Cycle IV: the matching upper bound, and exact semiprime splitting -/Formal statement
theorem HyperbolicBerggrenGeodesics.ball_card_upper{R : ℝ} (hR : 0 ≤ R) (S : Finset (ℕ × ℕ))
(hseed : ∀ p ∈ S, IsSeed p.1 p.2)
(hball : ∀ p ∈ S, ∀ h : 0 < p.1, dist (hpoint p.1 p.2 h) UpperHalfPlane.I ≤ R) :
(S.card : ℝ) ≤ 4 * Real.exp (2 * R) := by sorrySource