On the interpolation of integer-valued polynomials

From MaRDI portal



Abstract: It is well known, that if polynomial with rational coefficients of degree n takes integer values in points 0,1,...,n then it takes integer values in all integer points. Are there sets of n+1 points with the same property in other integral domains? We show that answer is negative for the ring of Gaussian integers mathbbZ[i] when n is large enough. Also we discuss the question about minimal possible size of set, such that if polynomial takes integer values in all points of this set then it is integer-valued.


Let \(R\) be an integral domain and let \(K\) be its field of fractions. A subset \(M\) of \(R\) is said to be \(n\)-universal if every polynomial \(f\in K[X]\) of degree \(n\), satisfying \(f(M)\subset R\) is integral-valued, i.e. one has \(f(R)\subset R\). It has been shown by \textit{M. M. Wood} [J. Number Theory 99, No. 1, 36--56 (2003; Zbl 1076.13011)] that if \(R\) is the ring of integers of an imaginary quadratic field, then there is no infinite sequence in \(R\) whose first \(n+1\) elements form an \(n\)-universal set. The authors strengthen this result in the case \(R=Z[i]\) by showing that for sufficiently large \(n\) there are no \(n\)-universal sets of cardinality \(n+1\), confirming a conjecture of \textit{D. Hensley} [J. Number Theory 9, 510--524 (1977; Zbl 0365.10035)]. They show also the existence in \(Z[i]\) of \(n\)-universal sets of cardinality \(O(n)\).











This page was built for publication: On the interpolation of integer-valued polynomials

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q740939)