Calculating characters of \(p\)-groups (Q1813914)

The author gives an algorithm for the computation of the characters of a \(p\)-group \(G\) given by a special power-commutator presentation for \(p\)- groups, called \(p\)-step generating series. This algorithm is a refinement of a former work of \textit{M. C. Slattery} [J. Symb. Comput. 2, 51-58 (1986; Zbl 0588.20006)]. It uses the fact that each irreducible character \(\chi\) of \(G\) is an induced character \(\lambda^ G\), where \(\lambda\) is a faithful linear character of a cyclic section \(C/K\) of \(G\), with \(K\triangleleft C\leq G\). The algorithm does not calculate all required sections and their irreducible characters and induces them up to \(G\) afterwards, but calculates the multiplicity with which each conjugacy class participates in these inductions by recursion. After this the linear characters of \(C/K\) can be computed and induced up to \(G\) easily. Thus it is possible just to print out the found character values instead of storing them into the workspace. Furthermore, an implementation of this algorithm in CAYLEY V3.6 is shown, and the benefits of different representations of the character values in the computer are pointed out. Finally, the author compares the run-time requirements of the resulting different implementations of this algorithm and the well known Dixon- Schneider Algorithm, that makes no use of any known structure of \(G\), for some groups.