Abels's groups revisited. (Q373643)

This paper explores topological and homological finiteness properties of a group \(G\). For a positive integer \(n\), a group \(G\) is said to be of type \(\text F_n\) if \(G\) admits a classifying space \(X\) whose \(n\)-skeleton is compact modulo the action of \(G\). On the algebraic side, \(G\) is said to be of type \(\text{FP}_n\) if the \(\mathbb ZG\)-module \(\mathbb Z\) admits a projective \(\mathbb ZG\)-resolution \(P_\bullet\to M\) such that \(P_k\) is finitely generated for \(k\leq n\). It is known that a group being of type \(\text F_n\) is necessarily of type \(\text{FP}_n\). The converse is not true, although \(G\) being of type \(\text{FP}_n\) and of type \(\text F_2\) implies that it is of type \(\text F_n\). Topologically, if one considers only proper actions on a classifying space, one gets the notion of \(G\) being of type \(\underline{\text{F}}_n\). On the algebraic side, with the use of Bredon modules and Bredon homology, there is a notion of \(G\) being of type \(\underline{\text{FP}}_n\), which is defined via the family of finite subgroups of \(G\) and an analogous condition on projective resolutions of the trivial Bredon module. One has similar implications between a group being of type \(\underline{\text{F}}_n\) and of type \(\underline{\text{FP}}_n\). Consider the algebraic group \(\text{GL}_{n+1}(\mathbb Z[1/p])\) for a prime \(p\) and the subgroup of upper triangular matrices with upper left and lower right entries being 1. These groups, introduced by Abels, have the property of being of type \(\text F_{n-1}\) but not of type \(\text F_n\). In this paper, the author constructs a family of upper triangular subgroups of \(\text{GL}_{n+1}(\mathbb Z[1/p])\) that generalize Abels's groups. For odd primes \(p\), these groups are shown to be of type \(\text F_{n-1}\) but not of type \(\text{FP}_n\) (and hence not of type \(\text F_n\)). Also, for \(0<m\leq n\), such groups can be constructed to be of type \(\underline{\text{F}}_{m-1}\) but not of type \(\underline{\text{FP}}_m\) (and hence not of type \(\underline{\text{F}}_m\)). To prove the desired properties, beginning from a Bruhat-Tits building associated to \(\text{GL}_{n+1}(\mathbb Q_p)\), the author constructs a space on which the group of interest acts cocompactly and for which the stabilizers under finite subgroups have certain connectivity properties. More specifically, the goal is to construct a space to which one can apply criteria of \textit{K. S. Brown} [J. Pure Appl. Algebra 44, 45-75 (1987; Zbl 0613.20033)] for the classical finiteness conditions. To obtain the claims about the Bredon finiteness conditions, the author uses an analogue of Brown's criterion recently obtained by \textit{M. G. Fluch} and \textit{S. Witzel} [Homology Homotopy Appl. 15, No. 2, 153-162 (2013; Zbl 1297.20053)].