Kleinian groups and the rank problem. (Q2388847)

The `rank problem' asks if there is an algorithm that, given a finite presentation of a group \(G\) from a specific class of finitely presented groups, computes the rank (the minimal number of generators) of \(G\). For example, for word-hyperbolic groups the word, conjugacy and (at least in the torsion-free case) isomorphism problems are known to be solvable whereas the rank problem is unsolvable. The authors proved in a previous paper that the rank problem is solvable for torsion-free, locally quasi-convex, word-hyperbolic groups. The major class of examples of word-hyperbolic groups comes from the world of Kleinian groups which are in general not quasi-convex. In the present paper, the authors consider the class \(\mathcal M\) of all torsion-free word-hyperbolic groups which admit a properly discontinuous action on hyperbolic space \(\mathbb{H}^3\). The main result of the paper states that the rank problem is solvable for groups \(G\) from the class \(\mathcal M\); also, every such group \(G\) has only finitely many Nielsen-equivalence classes of generating \(k\)-tuples, for every \(k\geq 1\). The main tools for the proofs are a generalization by the two authors of Nielsen's cancellation methods for groups acting by isometries on Gromov-hyperbolic spaces, and also the proof of the tameness conjecture for Kleinian groups due to Agol and, independently, Calegari and Gabai (which, together with work by Canary, yields a precise characterization of non-quasiconvex subgroups for Kleinian groups).