\(k\)-full integers between successive \(k\)-th powers (Q648951)

Let \(k\geq 2\) be a fixed integer and \(F_{k,m}(N)\) be the counting function of \(1\leq n\leq N\) for which there are \(m\) squarefull integers in the open interval \((n^k, (n+1)^k)\). (A squarefull number \(n\) is a natural number whose factorization in distinct prime powers \(n=p_1^{\alpha_1}\cdots p_r^{\alpha_r}\) has all exponents \(\alpha_i\geq 2\), \(i=1,\cdots,r\)). In the paper [Mathematika 27, 171--178 (1980; Zbl 0451.10027)] \textit{P. Shiu} applied a theorem of Cassels to show that the density \(F_{2,m}(N)/N\) tends to \(d_{2,m}\) as \(N\to \infty\) and gave an explicit formula for \(d_{2,m}\). In this paper the authors applied a different approach based on the Koksma-Hlawka inequality and Erdös-Turán-Koksma inequality. They showed that \[ \frac{F_{k,m}(N)}N= d_{k,m}+O_k((\log\log N)^{-1/(2k)}),\quad N\to \infty \] and the sequence \(\{d_{k,m}\}\) satisfies the generating function \[ \sum_{m=0}^\infty d_{k,m}z^m = \prod_{\lambda\in S_k} \big(1+(z-1)\lambda^{-1}\big) \] where \(S_k=\{n_1^{1+\frac1k}n_2^{1+\frac2k}\cdots n_{k-1}^{1+\frac{k-1}k}: \;n_1,\cdots, n_{k-1}\geq 2,\;\mu^2(n_1n_2\cdots n_{k-1})=1\}\).