Fixed conjugacy classes of normal subgroups and the \(k(GV)\)-problem. (Q854912)

Let \(G\) be a finite group. We let \(k(G)\) denote the number of conjugacy classes of \(G\). Suppose that \(V\) is a faithful \(F_pG\)-module. If \((|G|,|V|)=1\) then it has been recently shown that \(k(VG)\leq|V|\). This is based on the work of many authors (see the references in the paper) and depends upon the classification of finite simple groups. This is equivalent to an old question of Brauer about the number of characters in a block (known as the \(k(B)\) conjecture). One can reduce to the case where \(G\) acts irreducibly on \(V\). One of the most difficult cases to deal with (and which was involved in the last step) is the case when \(G\) does not act primitively on \(V\). The author considers a different approach to this question -- much more related to counting and internal group theory than some earlier methods. In particular, he gives a new proof of the last case handled in the solution of the \(k(GV)\) problem (Theorem 6.2). He also shows that for large enough primes, his methods give a reduction to the primitive case. An advantage of this approach is that it should prove very useful in studying the non-coprime \(k(GV)\) problem as well. One no longer assumes that \((|G|,|V|)=1\). The critical situation to study is when \(O_p(G)=1\) and \(G\) acts faithfully and irreducibly on \(V\). The result is no longer true in all cases (but in a certain sense the only known counterexample is when \(G=GL(V)\) and then it is still close to being true). Robinson has shown that there are some nice representation theory questions related to this. It has been proved to hold in some cases (Tiep and the reviewer showed that it almost always holds when \(G\) is close to being a simple group). Again, the reduction to the primitive case will be most difficult and Keller's results and methods will give results. In particular, his reduction theorem for large primes is still valid in this setting. The results and ideas in this paper will be quite useful in studying the non-coprime \(k(GV)\)-problem.