On the binary relations defined using GD1 and 1GD inverses over infinite dimensional vector spaces
Diego Alba Alonso* Javier Sánchez González** · 2026
Read On the binary relations defined using GD1 and 1GD inverses over infinite dimensional vector spaces by Diego Alba Alonso* Javier Sánchez González** (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-26723.txt.
Introduction
Alba and Sánchez study binary relations defined with two generalized inverses, GD1 and 1GD, for maps on vector spaces of infinite dimension. They work with finite potent endomorphisms, maps whose powers eventually stabilize, and describe both inverses through the AST decomposition of such a map. That description yields algorithms for computing the inverses and shows the two relations form partial orders, a ranking-like structure, on finite potent endomorphisms. The result completes the theory of these inverses beyond the finite matrix case.
Catalog record
| CREATOR | Diego Alba Alonso* Javier Sánchez González** |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 114381 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.26723
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On the binary relations defined using GD1 and 1GD inverses over infinite dimensional vector spaces Thanks: This work was supported by Agencia Estatal de Investigación (Spain) through grant PID2023-151823NB-I00. Diego Alba Alonso* Javier Sánchez González** Abstract . The purpose of this article is to study certain binary relations of endomorphisms over infinite dimensional vector spaces defined by GD1 and 1GD generalized inverses. In order to do so, these generalized inverses are studied over arbitrary vector spaces (namely, infinite dimensional ones) using finite potent endomorphisms. We characterize them in terms of the AST decomposition of a finite potent endomorphism and we obtain algorithms for their respective computation. This theory is then used to characterize the GD1 and 1GD binary relations for finite potent endomorphisms in terms of the AST decomposition and to prove that they define partial orders in the set of finite potent endomorphisms, thus, completing the theory of these generalized inverses for matrices. Mathematical Subject Classification : 06A06, 15A03, 15A04, 15A09. Keywords : Finite potent endomorphism, \{1\}- inverse, GD1 inverse, 1GD inverse, GD1 Partial Order, 1GD Partial Order. 1. Introduction. Let \operatorname{Mat}_{n\times s}(k) denote the ring of (n\times s) matrices over an arbitrary field k. Given A\in\operatorname{Mat}_{n\times s}(k), the index of A, i(A), is the smallest positive integer m such that \mathrm{rk(A^{m})}=\mathrm{rk(A^{m+1})}. We say that a matrix X\in\operatorname{Mat}_{s\times n}(k) is an inner inverse or \{1\}- inverse of A when A\cdot X\cdot A=A. These are a simple example of the so called generalized inverses of matrix A. In [ 17 ] , G-Drazin inverses of a matrix with entries in the complex numbers A\in\operatorname{Mat}_{n\times n}(\mathbb{C}) were introduced as the matrices X\in\operatorname{Mat}_{n\times n}(\mathbb{C}) such that A\cdot X\cdot A=A\quad\&\quad A\cdot X\cdot A^{m+1}=A^{m}\quad\&\quad A^{m+1}\cdot X=A^{m}, where m=i(A). It is customary to denote the set of \{1\}- inverses of a matrix A by A(1) , as well as using the notation A^{-}\in A(1) for its elements (although they are not unique). Similarly, A(GD) denotes the set of G-Drazin inverses of a matrix A and any of its elements are denoted by A^{GD} (even though they are not unique). These generalized inverses have been generalized to the context of Banach spaces, see [ 8 ] . Further, in [ 6 ] , a new generalized inverse was studied combining \{1\}- inverses and G-Drazin inverses. Precisely, the composition A^{GD}\cdot A\cdot A^{-} for certain A^{-}\in A(1) and A^{GD}\in A(GD) was called as GD1 inverse of A. Different GD1 inverses appear varying the \{1\}- inverse and the G-Drazin inverse considered. Simultaneously, 1GD inverses were introduced as the “duals” of GD1 inverse, this is, as the composition A^{-}\cdot A\cdot A^{GD}. Again, these generalized inverses were studied in the framework of Banach space operators in [ 15 ] . It is worth pointing out that in this work, GD1 and 1GD inverses were investigated using the core-quasinilpotent decomposition as well as the closed range decomposition operator. Moreover, the interconnection with the Drazin inverse (in whatever way one could define it, as we will not use it in this paper) and other properties considering idempotent conditions and projectors were analyzed. Our treatment of this topic is completely different from this point of view. In both works, [ 6 ] and [ 15 ] , a binary relation dealing with GD1 inverses (and 1GD inverses respectively) was introduced. In [ 6 , Theorem 2.14, Theorem 3.14] , these relations were proved to be equivalent to other known relations, giving as an immediate corollary that they were partial orders in the set of matrices of index lesser or equal than 1. In [ 16 ] , John Tate introduced the concept of finite potent endomorphisms. Let V denote an arbitrary k- vector space, in general, infinite-dimensional. The ring of endomorphisms over the vector space V will be denoted as \operatorname{End}_{k}(V). Given \varphi\in\operatorname{End}_{k}(V) we will say that \varphi is finite potent when \varphi^{n}(V) is a finite dimensional k- vector subspace of V, where \varphi^{n}=\varphi\circ\dots\circ\varphi composed n times. Further, in [ 4 ] , these endomorphisms were characterized in terms of a \varphi- invariant decomposition of the vector space which will be called AST decomposition. This article is a contribution to the theory of Generalized Inverses and Matrix Partial Orders. Namely, we generalize the theory of GD1 inverses and 1GD inverses to arbitrary vector spaces, in general, infinite dimensional ones, using the theory of finite potent endomorphisms. We extend the studied binary relations defined by these generalized inverses and we prove that they are partial orders in the set of finite potent endomorphisms over arbitrary vector spaces. The outline of this work, together with the results obtained, can be summarized as follows: • Generalization of GD1 (respectively 1GD) inverses to infinite dimensional vector spaces (without inner product, nor Banach or Hilbert spaces) in Section 3.2 (respectively Section 3.2 ). A characterization of these generalized inverses using the AST decomposition, Theorem 3.10 (respectively Theorem 3.22 ). Moreover, study of the structure of the set of all GD1 inverses of a finite potent endomorphism (respectively 1GD), Theorem 3.11 (Theorem 3.23 ). • Application of the previous theory to Matrix Theory. Bearing in mind the well known relationship between endomorphisms over finite dimensional vector spaces and finite square matrices, the previous theory is used to study the set of GD1 inverses of a finite square matrix over an arbitrary field (respectively the set of 1GD inverses of a finite square matrix). Namely, if A=A_{1}+A_{2} is the core-nilpotent decomposition of A (Section 2.0.2 and \mathrm{N_{u}(A)} denotes the nullspace of A, then by Theorem 3.12 and Theorem 3.24 , we have bijections: \displaystyle A(GD1)\simeq k^{[\mathrm{dim}N_{u}(A)]\cdot(\mathrm{rk(A)}+\mathrm{rk(A_{2})})} \displaystyle A(1GD)\simeq k^{[\mathrm{dim}N_{u}(A)]\cdot(\mathrm{rk(A)}+\mathrm{rk(A_{2})})}. Moreover, the algorithms for the explicit calculation of these sets are offered in Section 3.1.1 and Section 3.2.1 , as well as some illustrative examples, in Section 3.3 . • In Section 4.1 the binary relations defined by GD1 and 1GD inverses in [ 6 ] are extended and studied over infinite dimensional vector spaces. After characterizing them using the AST decomposition of the operators involved, in Theorem 4.18 and Theorem 4.30 , we prove that they are both partial orders in the set of finite potent endomorphisms (Theorems 4.20 and 4.31 ). This completes the theory of these binary relations, as in the matricial case presented in [ 6 ] , the authors prove that these relations are equivalent to other known ones defining partial orders (see [ 6 , Theorem 2.14, Theorem 3.14] ), but only in set of matrices of index lesser or equal than 1. Moreover, we give a non trivial example (Example 3 ) of matrices ordered for these relations which are of index different than 1, using the characterizations that we presented previously. • As a side note we show that there is no relation between the binary relations involving GD1 and 1GD inverses and the subsets of GD1 and 1GD inverses of the linear operators (or matrices) defining this relation, Remark 4.13 . • To conclude, we present two “new” binary relations using compositions of GD1 and 1GD inverses (Definitions 4.33 and 4.34 ) and we show that they are equivalent to the minus partial order and the G-Drazin partial order respectively (Theorem 4.37 ). We shall point out that as far as the authors