Regular orbits of linear groups with an application to the \(k(GV)\)-problem. I (Q1569804)

Let \(p\) be a prime and let \(F\) denote the field of order \(p\). Let \(G\) be a finite group of \(p'\)-order and let \(V\) be a faithful \(FG\)-module of finite dimension \(n\). A fundamental question that occurs frequently in group theory is the following: when does \(G\) have a regular orbit on \(V\)? It is extremely difficult to answer this question from first principles, and it has often been necessary to assume that \(G\) is solvable to obtain any progress. Even in the solvable case, it is not easy to decide if regular orbits exist when \(p\) is small. In the present paper, and its second (and concluding) part [ibid. 227, No. 2, 433-473 (2000; see the following review Zbl 0970.20006)], the author assumes that \(G\) contains a quasi-simple normal subgroup \(H\) that acts irreducibly on \(V\). Of course, to make any progress under this hypothesis, he must invoke the classification theorem for finite non-Abelian simple groups. The author's main theorem may be stated as follows (under the hypotheses described above). Suppose that \(G\) has no regular orbit on \(V\). Then either \(H\) is the alternating group \(A_m\), where \(m<p\), and \(V\) is the deleted permutation module for \(A_m\) (the deleted permutation module for \(A_m\) is a submodule of codimension 1 in the natural permutation module of dimension \(m\) over \(F\)), or else \(H/Z(H)\), \(n\) and \(p\) belong to a finite list of possibilities given in the paper (here \(Z(H)\) is the centre of \(H\)). The largest value of \(n\) occurring in the list is 12 and the largest value of \(p\) is 271. The possibilities for \(H/Z(H)\) are, in the usual notation: \(A_5\), \(A_6\), \(A_7\), \(L_2(7)\), \(L_3(4)\), \(U_3(3)\), \(U_4(2)\), \(U_4(3)\), \(U_5(2)\), \(\text{Sp}_6(2)\), \(\Omega^+_8(2)\), \(J_2\) and \(Suz\). From the list we can deduce that \(p\leq 23\), unless \(H/Z(H)\) is one of the five following possibilities: \(A_5\), \(U_4(2)\), \(U_4(3)\), \(\text{Sp}_6(2)\) and \(\Omega^+_8(2)\) (several of these groups occur in Weyl groups). The author makes various illustrative comments in his introduction. In the alternating group case, we can find an action in which no regular orbits exist. We take \(m=p-1\), \(n=p-2\) and \(G\) to be the direct product of the multiplicative group of \(F\) with the symmetric group \(S_{p-1}\). In the second case, genuine examples have been constructed when \(p\leq 61\). When \(p=61\), we may take \(n=2\) and \(G\) to be the group generated by \(\text{SL}_2(5)\) and the multiplicative group of \(F\). The author's work has an application to the \(k(GV)\)-problem, as the title suggests. To explain this application, suppose that \(H\) is a finite group and \(B\) is a \(p\)-block of irreducible characters of \(H\) with defect group \(D\). Brauer's celebrated \(k(B)\)-problem is to show that the number of irreducible complex characters of \(H\) in \(B\) is at most \(|D|\). To solve Brauer's \(k(B)\)-problem when \(H\) is \(p\)-solvable, it suffices to assume that \(H=VG\), where \(V\) is a minimal normal self-centralizing elementary Abelian \(p\)-subgroup of \(H\) and \(G\) is a \(p'\)-group. Then we must show that the number of conjugacy classes in \(H\) is at most \(|V|\). This has become known as the \(k(GV)\) problem. Considering \(V\) to be a faithful irreducible module for \(FG\), a simple corollary of Goodwin's main theorem implies that Brauer's \(k(B)\)-problem has an affirmative answer for all \(p\)-solvable groups \(H\) provided \(p\geq 211\). In this fast-moving subject, subsequent work by various authors has improved this result to take in all primes with the possible exception of \(3\), \(5\), \(7\), \(13\) and \(19\). The paper under review confines itself to investigating the case when \(H/Z(H)\) is either an alternating group or a sporadic simple group. A considerable body of specialized information about representations of finite simple groups is required to make progress. The author also builds on the work of his research supervisor, \textit{M. Liebeck}, [J. Algebra 184, No. 3, 1136-1142 (1996; Zbl 0901.20006)], whose bounds for \(p\) are considerably improved.