The density of sums of distinct divisors

Scott D. Hughes

En palabras de los autores

For a positive integer $t$ , let $d_{t}$ denote the natural density of the set of $n$ for which $t$ is a sum of distinct divisors of $n$ . Erd\H{o}s proved that $d_{t}$ exists, gave an unspecified polylogarithmic upper bound, asserted without proof a matching lower bound, and asked whether $d_{t} \sim c_{3} / (lo g t)^{c_{4}}$ . We record the explicit bounds \[ \nu_t \le d_t \ll \frac{(\log\log t)^{\delta-3/2}}{(\log t)^{\delta}}, \qquad \nu_t = \frac{K}{\log t}\Bigl(1 + O\Bigl(\frac{\log\log t}{\log t}\Bigr)\Bigr), \] where $δ = 0.086071 \dots$ is the Erd\H{o}s--Ford--Tenenbaum constant, $K = c e^{- γ}$ , and $c = 1.33607 \dots$ is the practical-number constant. Consequently, if Erd\H{o}s's asymptotic holds, then $δ < c_{4} \leq 1$ . A two-prime construction, using a half-scale sumset to obtain full residue coverage, then yields the pointwise excess \[ \liminf_{t\to\infty}(\log t) (d_t-\nu_t) \ge KI, \] where $I = \int_{0}^{2} G (w) d w = 0.05887 \dots$ is an explicit elementary integral. In particular $d_{t} \geq 0.79/ lo g t$ for every sufficiently large $t$ , and $d_{t}$ is not asymptotic to $K / lo g t$ .

Resultado principalEl resumen no menciona limitaciones.

The density of sums of distinct divisors

En palabras de los autores

Más de Álgebra y teoría de números ese día

The Odlyzko-Poonen Conjecture on Irreducibility of Random Polynomials

Solution to Dixmier's Problem about spectra of C*-algebras

Poincar\'e Duality and Quadratic Refinements over Laurent Rings

On Haagerup-Izumi fusion categories

Monoidal Gr\"obner systems and categories of affine Brauer type