Bennett exists integral root ratio bound
ProvedD1Bound5.bennett_exists_integral_root_ratio_bounddiophantine-tuplesnumber-theory
A uniform root-transfer estimate with the same final constant3 as the manuscript. The stronger entry bounds permit the simple factor3/2 here, avoiding a monotonicity argument for the auxiliary function e(X).
Preamble
import Lean.Elab.Tactic.Omega import Mathlib.Algebra.BigOperators.Fin import Mathlib.Algebra.BigOperators.NatAntidiagonal import Mathlib.Algebra.Polynomial.BigOperators import Mathlib.Algebra.Polynomial.Degree.Operations import Mathlib.Analysis.Analytic.Binomial import Mathlib.Analysis.Complex.Basic import Mathlib.Analysis.Normed.Ring.Basic import Mathlib.Analysis.SpecificLimits.Normed import Mathlib.Combinatorics.Enumerative.Catalan.Basic import Mathlib.Data.Finset.Insert import Mathlib.Data.Sym.Basic import Mathlib.Data.Sym.Card import Mathlib.LinearAlgebra.Lagrange import Mathlib.RingTheory.Binomial import Mathlib.RingTheory.MvPolynomial.Symmetric.Defs import Mathlib.RingTheory.PowerSeries.Binomial import Mathlib.Tactic.LinearCombination import Mathlib.Tactic.NormNum import Mathlib.Tactic.Ring
Formal statement
section
namespace D1Bound5
theorem _root_.D1Bound5.bennett_exists_integral_root_ratio_bound
{R : Type*} [CommRing R] (f : R →+* ℂ)
(a c d z r v : R)
(ha : f a ≠ 0) (hc : f c ≠ 0) (hz : f z ≠ 0)
(hr : r ^ 2 = a * c + 1) (hv : v ^ 2 = a * d + 1)
(hzsq : z ^ 2 = c * d + 1)
(hAC : 1000 ≤ ‖f a‖ * ‖f c‖) (hC : 4 * ‖f a‖ ≤ ‖f c‖) :
∃ s : R, (s = r * v ∨ s = -(r * v)) ∧
‖(1 + 1 / (f a * f c)) ^ (1 / 2 : ℂ) - f s / (f a * f z)‖ <
(3 / 2 : ℝ) * ‖f c‖ / (‖f a‖ * ‖f z‖ ^ 2) := by sorry
end D1Bound5
end
Source
D1Bound5/BennettRootApproximation.lean, lines 86-117, SHA-256 ca1061aebc87736a011fd6e5cb0011ac939b7397422bdbe21a20c1861b490805. Project: https://github.com/tomislav-math/d1-bound5-lean