Algorithms for arithmetic groups with the congruence subgroup property. (Q468700)

Thirty years back, Grunewald and Segal started the study of decision problems in arithmetic groups when they proved some beautiful, general results. In general, dealing with virtually solvable groups is easier and semisimple groups are rather hard to investigate. In recent times, the authors and many others like Bäärnhielm, Babai, Beals, O'Brien, Eick, de Graaf, Holt, Leedham-Green, Neunhöffer, Ostheimer, Rockmore and Seress have developed computational techniques to a good extent especially in solvable linear groups. This paper builds some computation techniques in arithmetic groups in a semisimple group like \(\mathrm{SL}(n,\mathbb Q)\). Determining whether a finitely generated matrix group is finite or, determining whether a finitely generated, infinite, solvable linear group is arithmetic are easier problems. However, the problem of determining whether a finitely generated subgroup of a group like \(\mathrm{SL}(n,\mathbb Z)\) is of finite index seems to be undecidable. However, if one knows that a certain subgroup is of finite index in this group, it is possible to use Coxeter-Moser-Todd type of algorithms to determine the index and other data. In particular, for \(n\geq 3\), the subgroups of finite index are congruence subgroups and the authors are able to determine the level computationally. One ingredient in these developments is a set of generators for principal congruence subgroups written down by the reviewer and Venkataramana. In principle, this can be carried out also in other arithmetic groups with the congruence subgroup property. The authors also need detailed computations in the finite groups \(\mathrm{GL}(n,\mathbb Z_m)\). All the algorithms are implemented in GAP by the authors.