🤖 AI Summary
This work addresses the computation of finite-index congruences on finitely presented semigroups and monoids. We propose two novel algorithms: one for infinite semigroups/monoids, generalizing Sims’ low-index subgroup algorithm; and another for finite cases, based on principal congruences and their unions. Our approach introduces, for the first time, the integration of relative Green’s relations and a monoid analogue of the Schreier lemma into the generation of principal congruences, significantly reducing computational redundancy. We develop the first open-source, cross-platform congruence computation toolkit—unified across GAP, C++ (libsemigroups), and Python—and integrate it into both the GAP Semigroups package and the main libsemigroups library. Experimental evaluation demonstrates that our method efficiently enumerates all finite-index left, right, and two-sided congruences on diverse classical semigroups, outperforming existing implementations in both speed and scalability.
📝 Abstract
In this paper we describe an algorithm for computing the left, right, or 2-sided congruences of a finitely presented semigroup or monoid with finitely many classes, and alternative algorithm when the finitely presented semigroup or monoid is finite. We compare the two algorithms presented to existing algorithms and implementations. The first algorithm is a generalization of Sims' low index subgroup algorithm for finding the congruences of a monoid. The second algorithm involves determining the distinct principal congruences, and then finding all of their possible joins. Variations of this algorithm have been suggested in numerous contexts by numerous authors. We show how to utilise the theory of relative Green's relations, and a version of Schreier's Lemma for monoids, to reduce the number of principal congruences that must be generated as the first step of this approach. Both of the algorithms described in this paper are implemented in the GAP package Semigroups, and the first algorithm is available in the C++ library libsemigroups and in its python bindings libsemigroups_pybind.