Finite soluble groups satisfying the swap conjecture. (Q902947)

Let \(G\) be a finite group and let \(d(G)\) be the minimal number of generators of \(G\). For any integer \(d\geq d(G)\), let \(V_d(G)=\{(g_1,\ldots,g_d)\in G^d\mid\langle g_1,\ldots,g_d\rangle=G\}\) be the set of all generating \(d\)-tuples. The ``swap graph'' \(\Delta_d(G)\) is considered: two vertices \((x_1,\ldots,x_d),(y_1,\ldots,y_d)\in V_d(G)\) are adjacent in the swap graph if and only if they differ only by one entry. \textit{R. F. Tennant} and \textit{E. C. Turner} [in Rocky Mt. J. Math. 22, No. 3, 1083-1095 (1992; Zbl 0790.20053)] proposed the conjecture that \(\Delta_d(G)\) is connected (``swap conjecture''). The author of the current paper and \textit{E. Crestani} proved [in J. Algebra 376, 79-88 (2013; Zbl 1287.20024)] that the swap conjecture is true if \(d\geq d(G)+1\), for an arbitrary finite group. For the more difficult case when \(d=d(G)\), [in Isr. J. Math. 198, 63-74 (2013; Zbl 1300.20021)], the same authors proved that the 2-generated finite soluble groups satisfy the swap conjecture, and they also analyzed a weaker version of the swap conjecture. In the paper under review it is proved that the swap conjecture is true if \(G\) is a finite group whose derived subgroup has odd order (Theorem 1), or it is nilpotent (Corollary 3).