Open rational interval tables certifying a strict finite theta Gram margin
OpenMasulloRH.cofinal_rational_theta_gram_interval_tablesFor arbitrarily large n, choose finitely many rational shifts and coefficients in Q+iQ for the dyadic theta translates, and set δ=1/[16(n+1)]. Form the augmented function list u consisting of the normalized Gaussian pulse followed by the translates, and the coefficient vector d=(1,−a₁,…,−aₘ). Let Γ be its endpoint-weighted Gram matrix at center 3/4 and half-width 1/4−δ.
The proposed input is a finite rational complex matrix Q and a rational error η≥0 such that every Gram entry is enclosed and the rational upper certificate has a strict margin:
The scalar certificate uses only rational arithmetic once the finitely many integral enclosures have been established. No inverse, positive-definiteness assertion, or zero-location premise is required. This is an OPEN cofinal family of numerical certificates, not a claim that a fixed finite table proves the infinite approximation assertion. The error allowance records how accurate the individual integral enclosures need to be for the chosen coefficients.
import Definitions.Def_MasulloRH_GaussianThetaApproximation open MasulloRH.GaussianApproximation MeasureTheory open scoped BigOperators ComplexConjugate set_option autoImplicit false
theorem MasulloRH.cofinal_rational_theta_gram_interval_tables :
∀ N : ℕ, ∃ n : ℕ, N ≤ n ∧ ∃ (m : ℕ) (a : Fin m → ℚ × ℚ) (y : Fin m → ℚ)
(R : Fin (m+1) → Fin (m+1) → ℚ × ℚ) (η : ℚ), 0 ≤ η ∧
let δ : ℝ := 1/(16*((n:ℝ)+1))
let u : Fin (m+1) → ℝ → ℂ := Fin.cons (pulse n)
(fun i x => theta (x-(y i:ℝ)))
let d : Fin (m+1) → ℂ := Fin.cons 1
(fun i => -(((a i).1:ℂ)+((a i).2:ℂ)*Complex.I))
let Q : Fin (m+1) → Fin (m+1) → ℂ :=
fun i j => ((R i j).1:ℂ)+((R i j).2:ℂ)*Complex.I
(∀ i j, ‖(∫ x : ℝ, (Real.exp (2*(3/4)*x+2*(1/4-δ)*|x|):ℂ) *
conj (u i x) * u j x) - Q i j‖ ≤ (η:ℝ)) ∧
(∑ i, ∑ j, conj (d i)*d j*Q i j).re +
(η:ℝ) * (∑ i, (|(d i).re|+|(d i).im|))^2 < δ/2 := by sorry