The Chebyshev bound behind Wiener--Ikehara #
For nonnegative coefficients a whose Dirichlet series has the Wiener--Ikehara boundary data
(summability on Re s > 1 and a remainder G = LSeries a - A / (s - 1) continuous on
Re s ≥ 1), the partial sums grow at most linearly:
∑_{1 ≤ n ≤ t} ‖a n‖ = O(t).
This is the Chebyshev-type bound that the Tauberian half of Wiener--Ikehara consumes to control
the tails of its test functions. It is derived from the boundary data and nonnegativity alone,
improving the crude O(t log t) bound TauCeti.LSeries.isBigO_sum_Icc_norm_of_boundary.
The argument has two steps.
- A window bound. Choose a smooth compactly supported test function
psiwhose Fourier transform is nonnegative, and at least somec > 0on a neighbourhood|v| < ηof the origin (TauCeti.exists_contDiff_hasCompactSupport_fourier_nonneg). The smoothed asymptoticTauCeti.LSeries.tendsto_tsum_term_mul_fourier_atTop_of_nonnegbounds the Fourier-weighted series∑ a n / n * 𝓕 psi (log (n / x) / 2π)for largex. All of its summands are nonnegative, and those withq x < n ≤ x, whereq = exp (-π η), carry a weight at leastc / x. Hence∑_{q x < n ≤ x} ‖a n‖ ≤ K x. - Summing the windows. A window bound with a fixed ratio
q < 1implies linear growth of the partial sums, by peeling off the window(q x, x]and recursing onq x(TauCeti.isBigO_sum_Icc_of_sum_Ioc_floor_mul_le).
Main results #
TauCeti.LSeries.isBigO_sum_Icc_norm_id_of_boundary: the Chebyshev bound∑_{1 ≤ n ≤ t} ‖a n‖ = O(t)for nonnegative coefficients with Wiener--Ikehara boundary data.
References #
- J. Korevaar, Tauberian Theory: A Century of Developments, Chapter III.
- G. Tenenbaum, Introduction to Analytic and Probabilistic Number Theory, Chapter II.
The Chebyshev bound. Nonnegative coefficients whose Dirichlet series is summable on
Re s > 1 and has a boundary remainder G = LSeries a - A / (s - 1) continuous on Re s ≥ 1
satisfy ∑_{1 ≤ n ≤ t} ‖a n‖ = O(t).