On a divisor problem related to the Epstein zeta-function (Q5917719): Difference between revisions

For \(k \geq 1\) a fixed integer let \(Z^k (s) = \sum^\infty_{n = 1} a_k (n)n^{-s}\) \((\text{Re} s > \ell/2),\) where \(Z(s)\) is the Epstein zeta-function of a positive definite quadratic form in \(\ell\) variables with integer coefficients. The author is interested in bounds for \(\Delta^*_k (x)\), the error term in the asymptotic formula for the summatory function of \(a_k (n)\). E. Landau's general results give \[ \Delta^*_k (x) \ll x^{c (k, \ell) + \varepsilon} \text{ with } c(k, \ell) = {\ell \over 2} - {\ell \over k \ell + 1} \text{ for fixed } k \geq 1,\;\ell \geq 2. \] The author uses Perron's inversion formula and bounds coming from the approximate functional equation for \(Z(s)\) to show that, for \(k \geq 2\) and \(\ell \geq 3\) fixed, \(c(k, \ell) = {\ell \over 2} - {1 \over k}\) is permissible.