scientific article; zbMATH DE number 2165668

From MaRDI portal
Publication:4675605





This paper investigates the question: For which values of \(k\) does the polynomial \(x^d+ 2\) take infinitely many \(k\)-free values? The best result until recently has been that of \textit{M. N. Huxley} and \textit{M. Nair} [Proc. Lond. Math. Soc., II. Ser. 41, 66--82 (1980; Zbl 0435.10026)], which shows that one can take \(k\geq d-2\) for \(d\geq 15\), \(k\geq d-3\) for \(d\geq 27\), and so on. This result applies to any irreducible polynomial without a fixed \(k\)th power divisor. The present paper gives a conditional treatment of the problem. Subject to a certain hypothesis about short incomplete exponential sums it is shown that one can take any \(k\geq{3\over 4}(d- 1)\) if \(d\) is even. Indeed one then gets the expected asymptotic formula for the number of \(k\)-free values of \(n^d+ 2\), both for unrestricted values of \(n\), and for prime values. Subject to an additional hypothesis on the size of solutions to the Pell equation, the bound on \(k\) is reduced to \((d- 2)\), thereby covering the case of square-free values of \(x^4+ 2\). In work in the course of publication the reviewer has given an unconditional treatment handling the range \(k\geq[(3d+ 2)/4]\) for general polynomials. The lower bound is worse than that of the present paper only by a bounded amount, but yields nothing new when \(d= 4\).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4675605)