On the non-coprime \(k(GV)\)-problem. (Q2438453): Difference between revisions

The \(k(GV)\)-theorem (formerly known as \(k(GV)\)-problem) is the well-known result that \(k(GV)\leq |V|=p^n\), where \(k(GV)\) is the number of conjugacy classes of the semidirect product \(GV=V\rtimes G\) for a finite-dimensional vector space \(V\) over a prime field \(\mathbb F_p\) and a finite \(p'\)-subgroup \(G\) of \(\mathrm{GL}(V)\) (see e.g. [\textit{P. Schmid}, The solution of the \(k(GV)\) problem. London: Imperial College Press (2007; Zbl 1153.20001)]). Since the completion of the proof attention now has turned to what has become known as the non-coprime \(k(GV)\)-problem which asks for strong bounds on \(k(GV)\) without the coprimeness hypothesis. The gist of the paper under review appears to be that while the bound \(k(GV)\leq |V|\) is no longer always true, it is still true ``very often'' and in some important situations, and the authors prove some nice results along these lines. For example, using variations of the Hall-Higman bounds on dimensions of eigenspaces, they treat the extraspecial case of the non-coprime \(k(GV)\)-problem. They also look at meta-cyclic groups. The third main result is the following. Theorem. Let \(V\) be a finite faithful primitive irreducible \(FG\)-module for some group \(G\) with \(F^*(G)=F(G)\). Then \(k(GV)\leq\max\{|V|,2^{1344}\}\). An upper bound on \(k(G)\) in terms of \(|V|\) is also included.