Abelian groups as autocommutator subgroups. (Q2255124)

Let \(G\) be a group and let \(\Aut(G)\) denote its automorphism group. For \(g\in G\) and \(a\in\Aut(G)\), the element \([g,a]=g^{-1}g^a\) is the autocommutator of \(g\) and \(a\). For a subset \(B\) of \(\Aut(G)\) one may then consider the subgroup \([G,B]\) of \(G\) generated by the autocommutators \([g,b]\) for \(g\in G\) and \(b\in B\). The authors consider an arbitrary finite abelian group \(A\) and they are asking whether the equation \([X,B]\cong A\) is solvable for \(X\) finite and abelian, where \(B\) is either \(\Aut(X)'\), or \(\Aut(X)^2=\{a^2\mid a\in\Aut(X)\}\), or \(\Aut(X)^{2k+1}=\{a^{2k+1}\mid a\in\Aut(X)\}\), or \(\mathrm{Inv}(\Aut(X))=\{a\in\Aut(X)\mid|a|=2\}\). It is shown that there are solutions for arbitrary \(A\) in the second and the last case and the precise structure of the corresponding autocommutator subgroups is determined in every case. The technical complications appearing in the particular case when \(X\) is abelian suggest that these problems become even more complex when trying to solve such equations for \(X\) an arbitrary finite group.