Normal coverings and pairwise generation of finite alternating and symmetric groups. (Q2437948): Difference between revisions

This paper considers several numerical constants associated to a group \(G\) which measure covering by proper subgroups and generation by pairs of elements. The main constant under consideration is the \textit{normal covering number} \(\gamma(G)\), which is the minimal number of proper subgroups of \(G\) such that every element of \(G\) is conjugate to an element of one of these subgroups. Related to this is a parameter \(\kappa(G)\) (which does not yet seem to have been named), which is the largest size of a set \(X\) of conjugacy classes of \(G\) such that any two elements from different classes in \(X\) generate \(G\). The main result of the paper is that there is a positive constant \(c\) such that if \(G_n\) is the symmetric or alternating group of degree \(n\), then \(\kappa(G_n)\geqslant cn\). Since it has already been established that \(\gamma(G)\leqslant 2n/3\) and it is known that \(\kappa(G)\leqslant\gamma(G)\) for any finite \(G\), this shows that both \(\kappa(G_n)\) and \(\gamma(G_n)\) grow linearly in \(n\). The paper is almost entirely devoted to the proof of the main theorem, which is very well presented. The proof is relatively elementary, involving some very fine calculation and estimation, and consideration (and enumeration) of cycle types which can occur for elements of the three types (primitive, imprimitive and intransitive) of maximal subgroups of symmetric groups. The paper ends with an actual estimate for the constant \(c\), in the case of the symmetric group of degree \(n\) with \(n\) even; here it is shown that (for sufficiently large \(n\)) one can take \(c=0.025\). Although this paper is very well written, it may not attract a large readership, for two reasons. Firstly, the techniques used in the proof are not particularly general, with few ideas that could be exported to other proofs. Secondly, the authors do very little to put the results in perspective -- the paper would benefit from an explanation of why the normal covering number is an interesting and useful parameter to consider, and what is known about the normal covering numbers of other (finite) groups.