On the function \(\sigma_ \delta(n)\) (Q1381638): Difference between revisions

Let \(A_\delta(x)= |\{n: \sigma_\delta(n)\leq x\}|\), where \(\delta>0\) and \(\sigma_\delta(n)= \sum_{d\mid n} d^\delta\). The author applies standard analytic techniques to prove that \[ A_\delta(x)= a_\delta x^{\frac{1}{\delta}}+ O_\delta (x^{\frac{1}{\delta}} \exp(-\lambda(x))), \] where \(\lambda(x)= (1+\frac{c\log_3x}{\log_2x}) (\frac{1}{2}\log x\log_2x)^{1/2}\), \(c=c(\delta)>0\) and the positive constant \(a_\delta\) is explicitly given. \textit{P. T. Bateman} [Acta Arith. 21, 329-345 (1972; Zbl 0217.31901)] obtained a slightly less precise result for the case \(\delta=1\). [Note: There are two misprints in the first few lines of the paper -- a reversed equality sign in the definition of \(A_\delta(x)\), and in the limit defining \(a_1\) in which ``\(x\to 0\)'' should be replaced by ``\(x\to\infty\)''].