Matrices with cyclically monotone rows and Cantor numeration systems
Pavel Šťovíček, Edita Pelantová · 2026
Read Matrices with cyclically monotone rows and Cantor numeration systems by Pavel Šťovíček, Edita Pelantová (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-2604-22344.txt.
Introduction
Šťovíček and Pelantová study square matrices with cyclically monotone rows, where each row runs down in a cycle from a maximal diagonal entry. They prove such a matrix is regular when every diagonal entry exceeds all other entries in its row. That property solves an open problem from Cantor numeration systems, systems for writing numbers in non-standard bases. The problem concerns a one-to-one link between alternate Cantor real bases and lists of p non-negative integer sequences meeting the Parry condition.
Catalog record
| CREATOR | Pavel Šťovíček, Edita Pelantová |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38292 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.22344
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Matrices with cyclically monotone rows and Cantor numeration systems Pavel Šťovíček, Edita Pelantová Abstract We study a class of square matrices with non-negative elements which have cyclically monotone rows in the sense that each row of a matrix from the class consists of a cyclically non-increasing sequence of numbers starting from a maximal element on the diagonal. We prove that if every diagonal element is strictly larger than all other elements in the respective row, then the matrix is regular. This property enables us to solve an open problem that comes from the theory of non-standard numeration systems, also called Cantor numeration systems. The problem concerns a one-to-one relationship between Cantor real bases, which are supposed to be alternate, that is, periodic with a period p p , and lists of p p sequences of non-negative integers satisfying the so-called Parry condition. Department of Mathematics, Faculty of Nuclear Science Czech Technical University in Prague Trojanova 13, 120 00 Praha, Czech Republic [email protected], [email protected] Keywords : non-negative matrices; cyclically monotone rows; Cantor real numeration system. MSC 15B48, 11A67, 11C20 1 Introduction Our main motivation in this paper is an open problem concerning the Cantor numeration systems. To solve it, we introduce and study a special class of square matrices whose rows are cyclically monotone. In more detail, each row of a square matrix from this class consists of a cyclically non–increasing sequence of non-negative numbers starting at the maximal element lying on the diagonal, see Section 2 for a formal definition. Theorem 13 , also in Section 2 , contains the main result of our article stating that the determinant of every matrix from this class is non-negative and, under a somewhat stricter assumption, the matrix is even regular. Let us note that the theorem may also be regarded as a partial generalization of a known result valid for circulant matrices [ 13 , Prop. 24] . Let us now briefly describe the nature and history of the above mentioned open problem. The history goes back to Cantor, who was the first to use positional number systems in which instead of one fixed integer \beta>1 chosen as a base, as customary in the decimal or binary system, a sequence \mathcal{B}=(\beta_{i})_{i\geq 0} of integers \beta_{i}>1 is employed. In this numeration system, the \mathcal{B} -expansion of a number x\in[0,1) - usually denoted (x)_{\mathcal{B}} - is a string a_{0}a_{1}a_{2}\cdots of integers a_{n} , 0\leq a_{n}<\beta_{n} , satisfying x=\sum_{n=0}^{+\infty}\frac{a_{n}}{\beta_{0}\beta_{1}\cdots\beta_{n}}\,. (1) Cantor focused on the problem of specifying conditions that guaranty the number defined by the series ( 1 ) to be irrational. It seems that Cantor’s original ideas had to wait almost a hundred years for further intensive development. In 1954, Cantor’s criterion for the irrationality of numbers defined by series ( 1 ) was generalized by Oppenheim [ 15 ] . Further generalizations have been later presented, for example, in papers by Hančl and Tijdeman, see [ 11 , 12 ] . Another problem related to the Cantor series ( 1 ) concerns statistical properties of the distribution of coefficients a_{n} . This topic was already addressed in the 1950s by Erdős and Rényi [ 9 ] , by Turán [ 20 ] , and later in the 1960s by Šalát [ 19 ] . Furthermore, the question of the normality of numbers in this numeral system has become the focus of recently published articles due to Airey and Mance and their collaborators, see [ 1 ] . The numeration system, as originally introduced by Cantor, is a frequent subject of research up to the present time, for instance, from the point of view of the theory of number approximations. In this regard, let us just mention a very recent paper [ 5 ] . While Cantor in [ 4 ] admitted only integer sequences \mathcal{B} , Caalim and Demegillo in [ 3 ] and, independently of them, Charlier and Cisternino in [ 6 ] , being inspired by the beta expansion of real numbers introduced by Rényi [ 18 ] , began to study real Cantor numeration systems . In such systems \mathcal{B} is a sequence of real numbers greater than one satisfying \prod_{n=0}^{+\infty}{\beta_{n}}=+\infty . Furthermore, condition 0\leq a_{n}<\beta_{n} is replaced by another one that requires a_{0}a_{1}a_{2}\cdots to be the lexicographically greatest string such that ( 1 ) holds (in the case of integer bases, both conditions are equivalent). \mathcal{B} is called a Cantor real base . If the sequence \mathcal{B} is purely periodic, the number system is called alternate . As shown in [ 3 ] and [ 6 ] , to an alternate base \mathcal{B} with period p one can assign p integer sequences {\bf d}^{(\ell)}=(d_{n}^{(\ell)})_{n\geq 0} , \ell=0,1,\ldots,p-1 , with the help of which it can be decided whether a string a_{1}a_{2}a_{3}\cdots represents some number x\in[0,1) . An important characteristic of sequences {\bf d}^{(\ell)} assigned to a base \mathcal{B} is that they satisfy the so-called Parry condition. However, conversely, the question remained unanswered until recently whether a base \mathcal{B} exists for any list of sequences that satisfies the Parry condition and, if so, whether it is uniquely given. In a recently published article [ 8 ] , the question of existence has been answered in the affirmative. Furthermore, uniqueness has also been proven, but in some particular cases only, namely, in [ 7 ] , in the case where every {\bf d}^{(\ell)} is eventually periodic and, in [ 8 ] , in the case where every sequence {\bf d}^{(\ell)}={d}^{(\ell)}_{0}{d}^{(\ell)}_{1}\cdots starts with a number {d}^{(\ell)}_{0}\geq 2 . In the present paper, we also contribute to the solution of this problem. In Sections 3 and 4 , with the aid of algebraic tools developed in Section 2 , we are able to prove the uniqueness of the base \mathcal{B} for every list {\bf d}^{(\ell)}=(d_{n}^{(\ell)})_{n\geq 0} , \ell=0,1,\ldots,p-1 , of sequences that satisfy the Parry condition. 2 Matrices with cyclically monotone rows Notation 1 . As common, \mathbb{Z}_{p} , p\in\mathbb{N} , stands for the cyclic group of order p . As a set, \mathbb{Z}_{p}:=\{0,1,\ldots,p-1\}, and the group operation in \mathbb{Z}_{p} is the modular addition. To simplify notation, whenever an addition j+k , with j,k\in\mathbb{Z}_{p} , occurs in any expression, the addition is understood as j+k\mod p . The symbol \mathbb{N} stands for the set of positive integers, no special symbol will be used for non-negative integers. In this section, we introduce and deal with a special class of square matrices. Definition 2 . A square matrix A=(a_{j,k})_{0\leq j,k<p},\ p\in\mathbb{N}, is said to have cyclically monotone rows if its entries are all non-negative and its rows are cyclically non–increasing in the following sense: \forall j\in\mathbb{Z}_{p},\forall k\in\mathbb{Z}_{p}\setminus\{p-1\},\ a_{j,j+k}\geq a_{j,j+k+1}. (2) If, moreover, either p=1\ \text{ and }\ a_{0,0}>0 (3) or p\geq 2\ \text{ and }\ a_{j,j}>a_{j,j+1}\ \ \text{for every\ }j\in\mathbb{Z}_{p} (4) we say that A is a matrix with strictly cyclically monotone rows . Example . Consider the matrices \left(\begin{matrix}5&4&3\\ 1&2&2\\ 2&2&3\\ \end{matrix}\right)\ \ \text{and}\ \ \left(\begin{matrix}5&4&4\\ 1&2&1\\ 2&2&3\\ \end{matrix}\right)\!. The former has cyclically monotone rows, but not strictly, while the latter one has strictly cyclically monotone rows. Remark 3 . Let us list some simple properties of matrices with (strictly) cyclically monotone rows that we will use in some proofs to follow. 1. Both