The non-coprime \(k(GV)\) problem. (Q2575689)

Let \(p\) be a prime, let \(H\) be a finite group with an elementary Abelian normal \(p\)-subgroup \(V\) such that \(C_H(V)=V\), and let \(G=H/V\). The paper under review is concerned with upper bounds for \(k(H)\), the number of conjugacy classes of \(H\). By the recent solution of the \(k(GV)\)-problem, it is known that \(k(H)\leq|V|\) in case \(|G|\) and \(|V|\) are coprime. Recently, G.~R.~Robinson has put forward conjectures for the non-coprime situation. The authors deal with the case where \(G\) is almost quasi-simple and \(O_p(G)=1\). They show that ``usually'' \(k(H)<|V|/2\). In the exceptional case \(k(H)\geq|V|/2\), let \(L\) denote the unique non-Abelian composition factor of \(G\). Then either \(L\) belongs to a specific list of 19 groups (and there are explicit bounds on \(|V|\) whenever \(G\) acts irreducibly on \(V\)), or \(L\) is a finite group of Lie type in characteristic \(p\). In the last case, suppose that \(G\) acts irreducibly on \(V\). Then either \(L\) is a classical group of untwisted Lie rank at most 6 (and there are explicit bounds on \(|V|\)), or \(M:=E(G)\) is a classical group, and \(V\) is closely related to the natural \(M\)-module.