Strongly scale-invariant virtually polycyclic groups (Q2102726)

A finitely generated group \(G\) is \textit{strongly scale-invariant} if there is an injective endomorphism \(\varphi:G\longrightarrow G\) such that the index \(|G:\varphi(G)|\) and the sub\-group~\(\bigcap_{n>0}\varphi^n(G)\) are both finite. It has been conjectured in [\textit{V. Nekrashevych} and \textit{G. Pete}, Groups Geom. Dyn. 5, No. 1, 139--167 (2011; Zbl 1236.20031)] that a finitely generated strongly scale-invariant group is virtually nilpotent. The aim of the paper under review is to partially answer this conjecture by proving that a strongly scale-invariant group is virtually polycyclic if and only if it is virtually nilpotent (Theorem 3.10). As a consequence of the methods involved in the proof, it is proved that, given a monomorphism \(\varphi\) of a virtually polycyclic group \(G\) having a finite Reidemeister number for all iterates \(\varphi^n\), then \(G\) is virtually nilpotent; here, the \textit{Reidemeister number} of \(\varphi\) is the number of equivalence classes of the relation \(x\sim y\iff\exists z\in G\,:\, x=zy\varphi(z)^{-1}\). Finally, the author shows that for a finitely generated virtually nilpotent group \(G\), the property of being strongly scale-invariant can be read off the unique radicable group associated with \(G\) (Theorem 4.1). This result implies for instance that being strongly scale-invariant for virtually nilpotent groups is preserved under taking finite index subgroups.