Truncated Perron inversion at half-integers for bounded coefficients
ProvedWeil.perron_halfInteger_error_of_bounded_coefficientsanalytic-number-theorydirichlet-series
Let be a complex sequence with , and write in its half-plane of absolute convergence. For every integer , set . There is a constant such that, for every such and every real ,
The half-integer cutoff keeps the jump of the Perron kernel away from all integer arguments. The formal statement uses the platform’s normalized truncated vertical integral.
Preamble
import Definitions.Def_Weil_perron import Mathlib.NumberTheory.LSeries.Basic
Formal statement
namespace Weil
theorem perron_halfInteger_error_of_bounded_coefficients
(a : ℕ → ℂ) (ha : ∀ n, ‖a n‖ ≤ 1) :
∃ C : ℝ, 0 ≤ C ∧ ∀ (N : ℕ), 1 ≤ N → ∀ (T : ℝ), 1 ≤ T →
let x : ℝ := N + 1 / 2
‖(∑ n ∈ Finset.Icc 1 N, a n) - perronIntegral (LSeries a) x 2 T‖ ≤ C * x ^ 2 / T := by sorry
end WeilSource
E. C. Titchmarsh, The Theory of the Riemann Zeta-function, second edition revised by D. R. Heath-Brown (1986), Lemma 3.12, equation (3.12.1), pp. 60–62, with s=0, c=2, bounded coefficients and half-integer x. See also the application in §14.25(C), p. 370. https://sites.math.rutgers.edu/~zeilberg/EM18/TitchmarshZeta.pdf