Abelian groups, Gauss periods, and normal bases (Q5927544)

A result on finite abelian groups is first proven: Theorem: Let \(G\) be any finite abelian group. Let \(S\) be a subset and \(K\) a subgroup of \(G\) such that \(G=\langle S,K\rangle\). Then for any direct product \(G = G_1 \otimes G_2 \otimes \cdots \otimes G_t\), there is a subgroup of the form \[ H = H_1 \otimes H_2 \otimes \cdots \otimes H_t , \quad H_i \triangleleft G_i ,\quad 1 \leq i \leq t, \] such that \(G = <S,H>\) and \(G/H \simeq G/K\). For any subgroup \(K \subset {\mathbb Z}_r^{\times}\) of order \(k\), a Gauss period of type \((n,K)\) is defined as \[ \alpha = \sum_{a \in K} g({\beta}^a), \] where \(\beta\) is a primitive \(r\)th root of unity in \({\mathbb F}_{q^{nk}}\) and \(g( \cdot)\) is a polynomial of a certain form. The following theorem is proved: Theorem: Let \(q=p^m\), where \(p\) is a prime. There exists an integer \(r\) such that a Gauss period of type \((n,K)\) is normal for \({\mathbb F}_{q^n}\) over \({\mathbb F}_q\) for some subgroup \(K\) of \({\mathbb Z}_r^{\times}\) iff \(\gcd (m,n) =1\), and if \(p=2\) then \(8\) does not divide \(n\). Another theorem is proved that characterizes when there is a subgroup \(K\) of order \(k\) in \({\mathbb Z}_r^{\times}\) such that \(\langle q,K\rangle= {\mathbb Z}_r^{\times}\). In addition results are given on the construction of low complexity normal bases.