On the smallest number of terms of vanishing sums of units in number fields (Q1786703): Difference between revisions

Let \(\mathfrak O\) be an order of an algebraic number field \(K\). For any natural number \(k \geq 3\) one asks whether there exist units \(\varepsilon_1, \dots, \varepsilon_k\) of \(\mathfrak O\) with \[ \varepsilon_1+ \dots + \varepsilon_k =0 \] with no proper vanishing subsum. Let \(\ell(\mathfrak O)\) (\(\ell_o(\mathfrak O)\), \(\ell_e(\mathfrak O)\), resp.) denote the smallest (odd, even, resp.) number \(k\geq3\), for which such units exist. If \(\mathfrak O\) equals the ring of integers of \(K\), write \(\ell(K)\) (\(\ell_o(K)\), \(\ell_e(K)\), resp.). The authors prove the following results: For any \(K\) with infinitely many units, \(\ell_e(K)\) (and therefore also \(\ell(K)\)) is finite. For any \(k \geq 3\) there exists an order (in a quadratic number field) with \(\ell(\mathfrak O)=k\) and (up to the exceptions \(k=4t^4-4t+2\)) a (imaginary, cubic) number field with \(\ell(K)=k\), but on the other hand, there exist only finitely many number fields \(K\) with unit rank one, for which \(\ell(K) \leq k\). The different behaviour of odd and even \(k\) is shown by the following results: For any \(d \geq 2\) there exist infinitely many number fields \(K\) of degree \(d\) with \(\ell_o(K)=\infty\), whereas for any \(d \geq 3\) there exist infinitely many number fields \(K\) of degree \(d\) with \(\ell_e(K)=4\). The ideas for the proofs are from classical algebraic number theory. Finally, an application to certain arithmetic graphs is given.