The characterization of finite elasticities. Factorization theory in Krull monoids via convex geometry (Q2170480)

The manuscript under review explains the connection between factorization invariants in block monoids (and thus in transfer Krull monoids) and convex geometry. Let \(G\) be a finitely generated Abelian group, and let \(G_0\) be a subset of \(G\). A finite sequence of elements of \(G_0\) is a zero-sum sequence if the sum of its elements is zero. The set of zero-sum sequences is a monoid with the concatenation operation, denoted \(\mathcal{B}(G_0)\). Minimal zero-sum sequences are those zero-sum sequences such that none of its proper sub-sequences are zero-sum sequences. Every zero-sum sequence can be expressed as a product (concatenation) of minimal zero-sum sequences, which are the atoms of \(\mathcal{B}(G_0)\). Atomic monoids in general are not factorial (every non-unit admits a unique factorization in terms of atoms), and so many invariants have been introduced in the literature to measure how wild the sets of factorizations of an element in an atomic monoid are. The length of a factorization is the number of atoms appearing in that factorization. The elasticity of a non-unit element is the supremum of the lengths of its factorizations divided by the minimum of the possible lengths. The elasticity of the monoid is the supremum of the elasticities of its non-unit elements. If \(G_0\) has finitely many elements, then zero-sum sequences translate to non-negative integer solutions of linear systems of equations with integer coefficients, with some of them in congruences. Thus in this setting the elasticity of \(G_0\) (as well as many other non-unique factorization invariants) is finite. This manuscript mainly deals with the case when \(G_0\) has infinitely many elements, starting with the case when \(G\) is torsion-free, and then studying the general (finitely generated) setting. It is not surprising that if we are dealing with positive combinations of elements in \(G_0\), which is itself a subset of \(\mathbb{Z}^d\) or of a lattice in \(\mathbb{R}^d\) (in the torsion-free setting), the connection with cones and convex geometry of \(\mathbb{R}^d\) comes into scene. We are somehow dealing with non-negative integer solutions to systems of linear equations, and if we focus on elementary atoms, with non-negative rational solutions (elementary atoms are those zero-sum sequences with minimal support, known in the literature with many other different names depending on the scope they have been introduced: strong atoms, extreme rays, or extreme integral generator). Actually, one of the first results that the author shows relating atoms in \(\mathcal{B}(G_0)\) and cones, states that the set of atoms of \(\mathcal{B}(G_0)\) is not empty if and only if 0 is in the interior of the cone spanned by \(G_0\). In the same line, one of the main results in the manuscript shows that \(\mathcal{B}(G_0)\) has finite elasticity if and only if \(0\) is not in the interior of the cone of \(G_0^\diamond\), where \(G_0^\diamond\) is the set of elements \(g\) in \(G_0\) for which there is a linearly independent subset \(X\subset G_0\) containing \(g\) and a sequence \(\{x_i\}_{i=1}^\infty\subseteq G_0\) of elements in the interior of the cone spanned by \(-X\) such that the coordinate of \(-x_i\) in terms of \(X\) corresponding to \(g\) tends to infinity (the author presents other characterizations of \(G_0^\diamond\)). The path to achieve this finite-elasticity result is hard and tedious. The author presents a bunch of tools and results related to them. First, the author gives results about filtered sequences, encasement and boundness. Then he recalls the definitions of positive bases and their connection with elementary atoms, introducing next the concept of Reay system (inspired in the characterization of positive bases given by Reay), which is later generalized to oriented Reay systems, and finally to virtual Reay systems (families of Reay systems whose limit are oriented Reay systems). These concepts are then used to introduce the idea of finitary sets and the set \(G_0^\diamond\). The authors develops a whole theory that enables to define \(G_0^\diamond\) that is then used to extend Lambert's result on the set of minimal non-zero non-negative solutions of a single linear Diophantine equation. This results in the definition of Lambert subsets of \(G_0\), their existence is equivalent to the fact that \(\mathcal{B}(G_0)\) has finite elasticity. Finiteness of the elasticity implies finiteness of the weak tame degree, sets of distances (Delta sets), the catenary degree and also some structure theorems on the sets of lengths and the support of the atoms of \(\mathcal{B}(G_0)\). The author then shows how to translate the results obtained for torsion-free finitely generated Abelian groups to finitely generated Abelian groups with torsion. In both cases, torsion and torsion-free, the author shows that the elasticity \(\rho(\mathcal{B}(G_0))\) is finite if and only if one if its refinements \(\rho_k(\mathcal{B}(G_0))\) is finite, and a bound for such potential \(k\) is given in terms of the rank of the group and the exponent of the torsion part of the group (for \(k\) a positive integer, \(\rho_k(\mathcal{B}(G_0))\) is the supremum of all the lengths of factorizations of elements that can be expressed as products of \(k\) atoms). A unit-cancellative commutative monoid \(H\) is a monoid such that the equality \(ua=a\), with \(u,a\in H\), implies that \(u\) is a unit in \(H\). A transfer Krull monoid \(H\) is a unit-cancellative monoid for which there exists an Abelian group \(G\), \(G_0\subset G\), and a monoid homomorphism \(\theta: H\to \mathcal{B}(G_0)\) such that (1) \(z\) is a unit if and only if \(\theta(z)\) is the trivial sequence, and (2) whenever \(\theta(z)\) factors as a product of atoms, there exists a factorization of \(z\) such that \(\theta\) maps its atomic factors into the atoms involved in the factorization of \(\theta(z)\). Sets of lengths and many other factorization invariants (including elasticity) of \(H\) are then governed by the corresponding invariants in \(\mathcal{B}(G_0)\). Thus, the results obtained for \(\mathcal{B}(G_0)\) are rephrased in the scope of transfer Krull monoids. Some parts of the book are quite technical. The lack of examples makes them difficult to follow. The whole development of the machinery needed to prove the main result is titanic.