A super-multiplicative inequality for the number of finite unlabeled arbitrary and T 0 T_{0} topologies

Ibtsam A. R. Alroily, Brahim Chaourar · 2025

Read A super-multiplicative inequality for the number of finite unlabeled arbitrary and T 0 T_{0} topologies by Ibtsam A. R. Alroily, Brahim Chaourar (2025) 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-2511-17263.txt.

Introduction

Ibtsam A R Alroily and Brahim Chaourar prove an inequality about counting topologies, the abstract structures that say which sets count as open on a finite set of points. Let f of n be the number of distinct such structures on n points, ignoring mere renaming of the points. The authors show that the count for a combined set is at least the product of the counts for its two parts.

The same product bound is proved for the labelled count, where points are kept distinct, and for the narrower class of T-zero topologies, in both labelled and unlabelled forms. Results of this shape matter because exact counts are known only for small sets. So bounds that link sizes give one of the few handles on how fast the numbers grow.

Catalog record

CREATORIbtsam A. R. Alroily, Brahim Chaourar
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE16 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-17263

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A super-multiplicative inequality for the number of finite unlabeled arbitrary and T 0 T_{0} topologies Ibtsam A. R. Alroily, Brahim Chaourar Abstract Let n n be a nonnegative integer, and f ⁡ ( n ) f(n) the number of unlabeled finite topologies on n n points. We prove that f ⁡ ( n + m ) ≥ f ⁡ ( n ) ​ f ​ ( m ) f(n+m)\geq f(n)f(m) both for the labeled and unlabeled cases. Moreover, we prove a similar inequality for labeled and unlabeled T 0 T_{0} topologies. 2010 Mathematics Subject Classification: Primary 05A20, 54B15; Secondary 05A15. Key words and phrases: finite topology; T_{0} topology; number of unlabeled topologies; inequalities; growth rate. 1 Introduction Finite topological spaces raise interesting combinatorial questions, most notably the number T(n) of distinct topologies on n points. Exhaustive enumeration has established T(n) for n\leq 18 [ 1 , 13 ] , yet a general formula remains far for reaching. The enumeration can be further refined by counting T(n,k) , the number of topologies on n points with k open sets. This remains a long-standing open problem, though some known cases exist. Key contributions come from Erné and Stege, who computed T(n,k) for n\leq 11 and arbitrary k in [ 5 ] , including related counts for T_{0} , connected topologies, and homeomorphism classes. Their results yielded all T(n,k) for k\leq 23 [ 6 ] . Asymptotic behavior is complex. Finding a simple closed-form expression or a straightforward asymptotic formula for T(n) has proven to be quite challenging. However, there are some known bounds and asymptotic estimates. For instance, it is known that T(n) grows faster than 2^{\binom{n}{2}} . However, very little is known about the growth behaviour of T(n) . Super-multiplicative inequalities involving T(n+m) , T(n) , and T(m) , for any nonnegative integers n and m , are also important for this question. In another field of combinatorics, matroid theory, a famous conjecture is that: f(n+m)\geq f(n)f(m) , where f(n) is the number of non-isomorphic matroids on n -element set [ 15 ] (see [ 11 , p. 594] ). It resisted three decades [ 3 , 9 ] . If we prove a similar inequality for the number of unlabeled finite topologies on n points, it would reveal some interesting structural properties about how topologies on disjoint sets combine to form topologies on their union. We can then get lower bounds on the growth of f(n) . For example, since f(2)=3 [ 13 ] , f(n)=f(2+(n-2))\geq f(2)f(n-2)=3f(n-2) , suggesting at least exponential growth with a base related to \sqrt{3} . Such inequalities could be a stepping stone towards a more precise understanding of the asymptotic growth rate. In addition of the analogy with matroid theory, this kind of inequality provides recursive approximation via maximum product. Super-multiplicativity enables recursive lower bounds: f(n)\geq\max_{1\leq k\leq n-1}\left\{f(k)\cdot f(n-k)\right\}. This formulation is useful for estimating f(n) when exact enumeration is difficult, and it reflects the idea that the most prolific decompositions dominate the growth. Let \tau_{i} , i=1,2 , be two topologies defined on finite and disjoint sets E_{i} . The direct sum \tau=\tau_{1}\oplus\tau_{2} is the collection \{O_{1}\cup O_{2} with (O_{1},O_{2})\in\tau_{1}\times\tau_{2}\} . A straightforward consequence is that the direct sum defines a topology on the disjoint union E_{1}\cup E_{2} . For the labeled case, it is not difficult to see that, by using the direct sum, we achieve our super-multiplicative inequality. However, in the unlabeled case, we cannot prove it by using the direct sum. Indeed, if n=m , then we get f(2n)\geq\frac{1}{2}[f^{2}(n)+f(n)] by using this naive operation because it is commutative: if \tau_{i} , i=1,2 , are two non-homeomorphic finite topologies, then \tau_{1}\oplus\tau_{2}=\tau_{2}\oplus\tau_{1} . This is far from what we need to prove in this case: f(2n)\geq f^{2}(n) . So, the direct sum fails here and a more elaborated (non-commutative) operation should be used for this purpose. It is what we call the w -sum. This operation is introduced in the coming section. Finite topologies have found concrete applications across a range of scientific and technological domains. In chemistry, they support the analysis of molecular graphs and topological indices used to predict chemical properties [ 10 ] . In image analysis, finite topological spaces provide a rigorous framework for modeling pixel connectivity and digital surfaces [ 7 ] . In automata theory, they aid in the classification and minimization of state-transition systems, with implications for control and automation [ 12 ] . Beyond these, finite topologies play a role in robotics, where they inform configuration space analysis and motion planning [ 8 ] ; in geographic information systems (GIS), where they model spatial relationships and adjacency structures [ 4 ] ; and in computer science, particularly in topological data analysis (TDA), where they contribute to the study of shape and connectivity in high-dimensional data [ 2 ] . These applications intersect with industrial sectors such as petrochemicals, electronics, automation, geospatial technology, and data science. The rest of the paper is organized as follows. In Section 2, we introduce the w -sum and prove some of its properties. In Section 3, we prove the main result. The last section is devoted to conclusions and further directions. 2 The w -sum operation Definition 2.1 . Let n be a nonnegative integer, E a finite set of cardinality n , and \tau a collection of subsets of E . We say that \tau is a topology on E , or on n points, if: (w1) \O,E\in\tau ; (w2) \tau is closed under union, that is, if A,B\in\tau , then A\cup B\in\tau ; (w3) \tau is closed under intersection, that is, if A,B\in\tau , then A\cap B\in\tau . Moreover, the size of \tau is n , E is called its ground set, and its members are called open sets. Let us recall what we mean by a topologies homeomorphism. Definition 2.2 . Let \tau_{i} , i=1,2 , be two topologies defined on finite sets E_{i} , respectively, and \varphi a mapping from E_{1} to E_{2} . We say that \varphi is an homeomorphism from \tau_{1} to \tau_{2} if: (1) \varphi is bijective; (2) Y=\varphi(X)\in\tau_{2} if and only if X\in\tau_{1} . In this case, we say that X and Y are homeomorphic, and \tau_{1} and \tau_{2} are homeomorphic. In general, two topologies are homeomorphic if such homeomorphism exists. The class of unlabeled topologies is the class of non-homeomorphic topologies. A direct consequence of the above definition is Proposition 2.3 . Two homeomorphic open sets have the same cardinality. Moreover, homeomorphisms keep inclusion (order), intersection, union, and complement. We denote by T(n) and f(n) the number of labeled and unlabeled topologies on n points, respectively. While the corresponding class for unlabeled topologies is denoted by \mathcal{T}(n) . Definition 2.4 . Let \tau be a finite topology defined on E , X\in\tau , and Y a finite set that is disjoint from E . (1) The intersection-topology \tau\cap X is the collection \{O\cap X for all O\in\tau\} . (2) The shift topology \tau-X is the collection \{O\backslash X for all O\in\tau\} . (3) The inverse shift topology \tau+Y is the collection \{O\cup Y for all O\in\tau\} . We recall the notion of co-topology. Definition 2.5 . Let \tau be a finite

Read the complete text