Constructible units in Abelian p-group rings (Q2639141)

This is a nice piece of research on the unit groups of abelian p-groups rings. It may prove to be important in number theory and for the computation of class numbers. In a way, it is gratifying to see some portion of the subject of group rings is still in good health. Let A be a finite abelian p-group and let S(A) be the group of units of ZA which have augmentation 1 (are normalized) and are invariant under the involution which sends each element of A into its inverse. It is well known that the units of ZA are given as a direct product \(T\times S(A)\) where the torsion T is just \(\pm A.\) One important subgroup of S(A) is \(\Omega\) (A), the group of circular (or cyclotomic) units (i.e. those which are cyclotomic - in the sense of number theory - in every Wedderburn component of ZA). The index of \(\Omega\) (A) in S(A) is finite and related to the ``real'' class number of ZA. If Vandiver's Conjecture [see \textit{L. Washington}, ``Introduction to cyclotomic fields'' (1982; Zbl 0484.12001)] is true, this index is relatively prime to p. Another subgroup of the normalized, symmetric units is introduced in this paper. It is \(\Lambda\) (A), group of constructible units, the subgroup of \(\Omega\) (A) generated by units of the form \(w_{c,b}=x^{\epsilon (c)}\cdot f_ c(x)\cdot x^{\epsilon (b)}\cdot f_ b(x)-kf_ q(x)\), where \(q=p^ m\) is the exponent for A, where \(\epsilon (c)=(1-c)/2 modulo q\) and where \(f_ c(x)=1+x+...+x^{c-1}\). (These units are called constructible.) One of the main results is that \(\Lambda (A)=\Omega (A)\) if and only if p is a regular prime and that, in general, the index of \(\Lambda\) (A) in \(\Omega\) (A) is a power of p.