On the number of topologies on a finite set

We denote the number of distinct topologies which can be defined on a set X with n elements by T(n). Similarly, T0(n) denotes the number of distinct T₀ topologies on the set X. In the present paper, we prove that for any prime p, T(pᵏ) ≡ k + 1 (mod p), and that for each natural number n there exists...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Algebra and Discrete Mathematics
Datum:2019
1. Verfasser: Kizmaz, M.Y.
Format: Artikel
Sprache:Englisch
Veröffentlicht: Інститут прикладної математики і механіки НАН України 2019
Online Zugang:https://nasplib.isofts.kiev.ua/handle/123456789/188421
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Zitieren:On the number of topologies on a finite set / M.Y. Kizmaz // Algebra and Discrete Mathematics. — 2019. — Vol. 27, № 1. — С. 50–57. — Бібліогр.: 8 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
_version_ 1859937058909323264
author Kizmaz, M.Y.
author_facet Kizmaz, M.Y.
citation_txt On the number of topologies on a finite set / M.Y. Kizmaz // Algebra and Discrete Mathematics. — 2019. — Vol. 27, № 1. — С. 50–57. — Бібліогр.: 8 назв. — англ.
collection DSpace DC
container_title Algebra and Discrete Mathematics
description We denote the number of distinct topologies which can be defined on a set X with n elements by T(n). Similarly, T0(n) denotes the number of distinct T₀ topologies on the set X. In the present paper, we prove that for any prime p, T(pᵏ) ≡ k + 1 (mod p), and that for each natural number n there exists a unique k such that T(p + n) ≡ k (mod p). We calculate k for n = 0, 1, 2, 3, 4. We give an alternative proof for a result of Z. I. Borevich to the effect that T₀(p + n) ≡ T₀(n + 1) (mod p).
first_indexed 2025-12-07T16:09:39Z
format Article
fulltext “adm-n1” — 2019/3/22 — 12:03 — page 50 — #58 Algebra and Discrete Mathematics RESEARCH ARTICLE Volume 27 (2019). Number 1, pp. 50–57 c© Journal “Algebra and Discrete Mathematics” On the number of topologies on a finite set M. Yasir Kizmaz Communicated by V. Lyubashenko Abstract. We denote the number of distinct topologies which can be defined on a set X with n elements by T (n). Similarly, T0(n) denotes the number of distinct T0 topologies on the set X. In the present paper, we prove that for any prime p, T (pk) ≡ k + 1 (mod p), and that for each natural number n there exists a unique k such that T (p+ n) ≡ k (mod p). We calculate k for n = 0, 1, 2, 3, 4. We give an alternative proof for a result of Z. I. Borevich to the effect that T0(p+ n) ≡ T0(n+ 1) (mod p). 1. Introduction Given a finite set X with n elements, let T(X) and T0(X) be the family of all topologies on X and the family of all T0 topologies on X, respectively. We denote the cardinality of T(X) by T (n) and the cardinality of T0(X) by T0(n). There is no simple formula giving T (n) and T0(n). Calculation of these sequences by hand becomes very hard for n > 4. The online encyclopedia of N. J. A. Sloane [1] gives the values of T (n) and T0(n) for n 6 18. For a more detailed discussion of results in the literature, we refer to the article by Borevich [4]. In the present paper, we prove that for any prime p, T (pk) ≡ k + 1 (mod p), and that for each natural number n there exists a unique k such that T (p+ n) ≡ k (mod p). We calculate k for n = 0, 1, 2, 3, 4. We give an alternative proof for a result of Z. I. Borevich [5] to the effect that T0(p+ n) ≡ T0(n+ 1) (mod p). Our results depend on basic properties of group action, which is the first time used in this problem as far as the author is aware. 2010 MSC: Primary 11B50, Secondary 11B05. Key words and phrases: topology, finite sets, T0 topology. “adm-n1” — 2019/3/22 — 12:03 — page 51 — #59 M. Y. Kizmaz 51 2. Main results Let G be a group acting on the finite set X. Then the action of G on X can be extended to the action of G on T(X) by setting gτ = {gU | U ∈ τ} where τ ∈ T(X) and gU = {gu | u ∈ U}. Now set Fix(T(X)) = {τ ∈ T(X) | gτ = τ for all g ∈ G}. Notice that if G is a p-group, |T(X)| ≡ |Fix(T(X))| (mod p) as every non-fixed element of T(X) has orbit of length a positive power of p. For a given topology τ ∈ T(X) and x ∈ X, we denote the intersection of all open sets of τ including x by Ox. Note that Ox ∈ τ as we are in the finite case. Definition 1. A base B of a topology τ is called a minimal base if any base of the topology contains B. Example. Let X = {a, b, c} be a set and let τ be a topology on X with τ = {∅, {a}, {a, b}, {a, c}, {a, b, c}}. Now if B is an arbitrary base for τ then we have U = ⋃ x∈U Bx where U ∈ τ and Bx ∈ B such that x ∈ Bx ⊆ U . Thus, we observe that B′ = {{a}, {a, b}, {a, c}} ⊆ B. Moreover, B′ is also a base for τ which implies that B′ is a minimal base for τ . Proposition 1. Let τ ∈ T(X) and Mτ be the set of all distinct Ox for x ∈ X. A base B of τ is a minimal base if and only if B = Mτ . By the proposition, we extend the definition: a base B on X is minimal if B = Mτ where τ is the topology generated by B. Lemma 1. τ ∈ Fix(T(X)) if and only if Mτ is G-invariant. Proof. Let τ ∈ Fix(T(X)) and Ox ∈ Mτ . We need to show that gOx ∈ Mτ . Let gx = y for g ∈ G. As gτ = τ , gOx ∈ τ . Thus, gOx is an open set containing the element gx = y. Hence, Oy ⊆ gOx. Since g−1y = x, we can show that Ox ⊆ g−1Oy which forces gOx = Oy. So, gOx ∈ Mτ ∀g ∈ G. Now assume that Mτ is G-invariant. For each U ∈ τ , we have U = ⋃ x∈U Ox. Then gU ∈ τ as gOx ∈ Mτ for all g ∈ G, and hence τ ∈ Fix(T(X)). Theorem 1. Let p be a prime and let k be a natural number. Then T (pk) ≡ k + 1 (mod p). Proof. Without loss of generality, let X be the cyclic group of order pk, that is, X = Cpk . Clearly X acts on X by left multiplication. By extending “adm-n1” — 2019/3/22 — 12:03 — page 52 — #60 52 On the number of topologies on a finite set this action, X acts on T(X). Notice that |T(X)| ≡ |Fix(T(X))| (mod p) as X is a p-group. It is left to show that Fix(T(X)) has k + 1 elements. Let τ ∈ Fix(T(X)) and let Ox, Oy ∈ Mτ for x, y ∈ X. Then (yx−1)Ox is an open set including y. Hence,Oy ⊆ (yx−1)Ox which means |Oy| 6 |Ox|. The other inclusion can be done similarly so |Ox| = |Oy| for all x, y ∈ X. Now if m ∈ Ox ∩ Oy, then Om ⊆ Ox ∩ Oy. As their orders are equal, we must have Ox = Oy or Ox ∩ Oy = ∅. Thus, X is a disjoint union of elements of Mτ , that is, X = ⋃·Ox∈Mτ Ox. It follows that |X| = |Mτ ||Ox| for any x ∈ X. Note that X also acts on Mτ by Lemma 1. Let e be the identity element of X and Stab(Oe) be the stabilizer of Oe in the action of X on Mτ . Then we obtain that |X : Stab(Oe)| = |Mτ | as the action is transitive. It follows that | Stab(Oe)| = |Oe| as |X| = |Mτ ||Oe| also holds. Since our action is induced from left multiplication in the group X, the product Stab(Oe)Oe is exactly set multiplication in the group X. Then we obtain that Oe = Stab(Oe)Oe ⊇ Stab(Oe)e = Stab(Oe). But the equality |Oe| = | Stab(Oe)| forces that Oe = Stab(Oe). Hence, the set Mτ is the set of left cosets of a subgroup of X. Since the chosen topology τ from Fix(T(X)) uniquely determines Mτ and Mτ uniquely determines a subgroup Oe of X, we have an injection from Fix(T(X)) to the set of all subgroups of X. Conversely, for a subgroup H of X, the set of left cosets of H forms an X-invariant minimal base for a topology. The topology τ generated by this base is an element of Fix(T(X)) by Lemma 1. Hence, the cardinality of Fix(T(X)) is equal to the number of the subgroups of X, which is k + 1. By applying the same method in the above proof, we can also show that T0(p k) ≡ 1 (mod p). Actually, Z. I. Borevich proved a more general result about T0(n). Now we establish an alternative proof for the theorem of Z. I. Borevich. Theorem 2 (Z. I. Borevich). Let p be a prime. If k ≡ l (mod p−1), then T0(k) ≡ T0(l) (mod p). Proof. It is equivalent to show that T0(n+ p) ≡ T0(n+ 1) (mod p) for an integer n > 0. Let C = Cp be the cyclic group of order p and N be a set with n elements. Without loss of generality, let X be the disjoint union of C and N so that the cardinality of X is p+ n. We define the action of C on X in the following way: for c ∈ C and for x ∈ X, c ∗ x = cx if x ∈ C and c ∗ x = x if x ∈ N . Then the action of C on X can be extended to “adm-n1” — 2019/3/22 — 12:03 — page 53 — #61 M. Y. Kizmaz 53 the action of C on T0(X). As |T0(X)| ≡ |Fix(T0(X))| (mod p), it is left to show that |Fix(T0(X))| = T0(n+ 1). Let τ ∈ Fix(T0(X)). Pick x, y ∈ C and a ∈ N where x 6= y. We know that yx−1Ox = Oy. Then we can observe that Ox ∩N = Oy ∩N as N is fixed by C. Similarly, Ox∩C and Oy ∩C are disjoint or equal. But we can not have Ox = Oy as it is a T0 topology. Hence, (Ox ∩C)∩ (Oy ∩C) = ∅. Then we obtain that Ox ∩C = {x}, and hence Ox = {x} ∪ S for a subset S of N . On the other hand, x(C ∩Oa) = C ∩Oa, which forces that C ∩Oa is either C or ∅, and hence Oa is either C ∪ T or T for a subset T of N . Set Y = {x} ∪ N then for each τ ∈ Fix(T0(X)), we have subspace topology τY = {U ∩ Y | U ∈ τ} on Y . Note that τY is a T0 topology as it is the subspace topology on Y induced from τ . Now our aim is to show that the mapping τ 7→ τY is a bijection from Fix(T0(X)) to T0(Y ). Let π, τ ∈ Fix(T0(X)) such that τY = πY . Then we have Ot ∩ Y = O′ t ∩ Y for all t ∈ Y where Ot ∈ Mτ and O′ t ∈ Mπ. Note that Ox ⊆ Y by the properties deduced in previous paragraph, and hence we obtain Ox = O′ x. We also have the equality Oa ∩Y = O′ a ∩Y for a ∈ N , which forces that Oa ∩ C = O′ a ∩ C is either C or ∅ by previous paragraph. Since Oa ∩ N = O′ a ∩ N , we get Oa = O′ a for a ∈ N . Then it follows that Ot = O′ t for all t ∈ Y . Since Oz = zx−1Ox for z ∈ C, we obtain that Ot = O′ t for all t ∈ X, which implies the equality τ = π. Thus, our map is one to one. Now, let π ∈ T0(Y ) and let Mπ be the minimal base of π. Set Ot =      tx−1O′ x if t ∈ C O′ t if t ∈ N and x /∈ O′ t C ∪O′ t if t ∈ N and x ∈ O′ t for each O′ t ∈ Mπ. We need to show that B = {Ot | t ∈ X} is a minimal base and the topology τ generated by B is a T0 topology. To show that it is a minimal base, we need to show that if a ∈ Ot then Oa ⊆ Ot for any a, t ∈ X. In fact, we will observe that Oa ⊂ Ot which shows that τ is a T0 topology. It can be done case by case but here we present only nontrivial cases. Fix a ∈ N and let a ∈ Ox. Note that a ∈ O′ x as O′ x = Ox, and hence we obtain O′ a ⊆ O′ x. If x ∈ O′ a then we have O′ a = O′ x which is not possible as π is a T0 topology. Thus, x /∈ O′ a which implies Oa = O′ a by our setting. Then we obtain that Oa ⊂ Ox. Now assume that x ∈ Oa. Then we get that x ∈ O′ a otherwise x /∈ Oa by our setting. Thus, we obtain O′ x ⊂ O′ a in a similar way. It follows that Ox ⊂ Oa as Ox = O′ x. As a result τ ∈ Fix(T0(X)) as B is a C-invariant minimal base. Moreover, “adm-n1” — 2019/3/22 — 12:03 — page 54 — #62 54 On the number of topologies on a finite set the equality τY = π holds, which concludes that the mapping τ 7→ τY from Fix(T0(X)) to T0(Y ) is a bijection, which completes the proof. Corollary 1. T0(p k) ≡ 1 (mod p) where k is a natural number and p is a prime number. The proof of next theorem is similar to the proof of Theorem 2 in the sense that both use the same technique. For clarity, we repeat some arguments. Theorem 3. For each natural number n, there exists a unique integer k such that T (p+ n) ≡ k (mod p) for all primes p. Proof. If n = 0, T (p) ≡ 2 (mod p) by Theorem 1. Now we can assume that n > 0. Let C = Cp be the cyclic group of order p and N be a set with n elements. We define the action of C on X as in the proof the previous theorem. Then the action of C on X can be extended to the action of C on T(X). As |T(X)| ≡ |Fix(T(X))| (mod p), it is left to show that |Fix(T(X))| does not depend on the choice of prime p. Due to Lemma 1, |Fix(T(X))| is equal to the number of C-invariant minimal bases on X. Let x, y ∈ C. Notice that Ox completely determines Oy as (yx−1)Ox = Oy. Then |Ox∩C| = |Oy∩C| as yx−1(Ox∩C) = Oy∩C. Now it is easy to see that Ox ∩ C and Oy ∩ C are either disjoint or equal. As cardinality of C is prime, Ox ∩ C is either {x} or C. If a ∈ N then we must have gOa = Oa for all g ∈ C as ga = a. Hence, Oa ∩ C is either ∅ or C. Thus, the elements of C have no contribution to the number of possible minimal bases. Now we know that such k exists. If k′ is also such an integer then k ≡ k′ (mod p) for all primes p. Hence k − k′ is divisible by all primes which forces k − k′ = 0. By the previous theorem, we see that k is uniquely determined by n, which leads the following definition. Definition 2. We define the integer sequence k(n) for n > 0 to be the unique number such that k(n) ≡ T (p+ n) (mod p) for all primes p. The sequence k(n) has appeared in the online encyclopedia of N. J. A. Sloane [2] with reference number (A265042) after this article appeared in arxiv. The proof of the theorem also gives an algorithm to calculate k(n) for a given n. Corollary 2. T (p+ 1) ≡ 7 (mod p) for all primes p, that is, k(1) = 7. “adm-n1” — 2019/3/22 — 12:03 — page 55 — #63 M. Y. Kizmaz 55 Proof. By following the proof the Theorem 3, it can be easily counted that there are exactly seven C-invariant minimal bases. Then the result follows. However, it is difficult to count C-invariant minimal bases for larger n to determine k(n). We develop a new method to calculate k(n) for larger n. But the method requires knowing values of T (s) for some s. Theorem 4. The sequence k(n) satisfies the following inequality T (n+ 1) + T0(n+ 1) 6 k(n) < 2T (n+ 1) for n > 1. Proof. We follow the proof of Theorem 3. We have k(n) = |Fix(T(X))| where |X| = n + p and |C| = p. Thus, we need to count C-invariant minimal bases. According to the proof, we have two main cases. Case 1: Ox ∩ C = C for x ∈ C. Then Oy ∩ C = C for all y ∈ C and Ox = Oy for all x, y ∈ C. If a ∈ N then Oa∩C is either C or ∅. Hence we can see the whole C as one element, that is, pass to the quotient X = X/ ∼ where x ∼ y if x, y ∈ C then there is a one to one correspondence between C-invariant topologies and quotient topologies. It follows that we have exactly T (n+ 1) possible sub-cases. Case 2: Ox ∩ C = {x} for x ∈ C. Let x, y ∈ C then when we define Ox, Oy is completely determined as Oy = yx−1Ox. Moreover, we have Ox ∩ N = Oy ∩ N as C acts trivially on N . We also have Oa ∩ C is either C or ∅. Now set Y = N ∪ {x} then the mapping τ → τY is one to one from Fix(T(X)) to T(Y ). Hence we can have at most T (n + 1) possible sub-cases. But we can not have exactly T (n + 1) possible sub- cases as the given map is not onto. To see this, set O′ x = {x, a} for a ∈ N and O′ a = {x, a} for a topology π ∈ T(Y ). Let τ ∈ Fix(T(X)) such that τY = π. Note that O′ a = Oa ∩ Y and Oa must be fixed by the action of C. Thus, we have C ∪ {a} ⊆ Oa. Since Ox = O′ x = {x, a}, we obtain that Oa ⊆ {x, a}. Then we obtain that Oa = {a} as it must be fixed by C, which is a contradiction. Thus, the map is not onto. Then we get that k(n) < 2T (n + 1). Moreover, the mapping is onto from Fix(T0(X)) ⊆ Fix(T(X)) to T0(Y ) ⊆ T(Y ).(See the proof of Theorem 2 for details of how the map τ → τY is one to one. It also shows why this map is onto when the target set is T0(Y ).) Thus, we have at least T0(n+1) sub-cases, and hence T (n+ 1) + T0(n+ 1) 6 k(n) < 2T (n+ 1). “adm-n1” — 2019/3/22 — 12:03 — page 56 — #64 56 On the number of topologies on a finite set Corollary 3. lim n→∞ k(n) T (n+ 1) = 2. Proof. We have 1 + T0(n+1) T (n+1) 6 k(n) T (n+1) < 2 by Theorem 4. In [6], it is proved that limn→∞ T0(n) T (n) = 1, and hence the result follows. Theorem 5. The sequence k(n) = 2, 7, 51, 634, 12623 for n = 0, 1, 2, 3, 4 respectively. Proof. If n = 0, the result follows from Theorem 1. Now assume that n > 1. We only show the calculation of k(2) and rest of them follow in a similar way. By previous theorem, we have T (3) + T0(3) 6 k(2) < 2T (3) so 48 6 k(2) < 58. Clearly, we have T (5) ≡ k(2) (mod 3) T (7) ≡ k(2) (mod 5) It follows that k(2) ≡ 6 (mod 15) by solving the above congruence relation. We obtain that k(2) = 51 as 48 6 k(2) < 58. For n = 3, 4, we have the same procedure. We should note that for n > 5, there is no unique solution satisfying the inequality. For example, k(5) ∈ {357593, 387623, 417653}. The closed form of k(n) seems to be another open problem. Hence calculation of k(n) for specific n or some better lower and upper bounds can be seen as new problems arising from this article. Acknowledgement The author would like to thank an anonymous referee for many in- valuable suggestions on the original manuscript. References [1] N. J. A. Sloane, Online Encyclopedia of Integer Sequences(Concerned with se- quences A000798,A001035 ) [2] N. J. A. Sloane, Online Encyclopedia of Integer Sequences(A265042) [3] M. Benoumhani, The number of topologies on a finite set, J. Integer Seq. 9 (2006), no. 2, Article 06.2.6, 9 pp. [4] Z. I. Borevich, Periodicity Of Residues of The number Of Finite Labeled Topologies, Journal of Soviet Mathematics, 24, No. 4, 391-395 (1984) “adm-n1” — 2019/3/22 — 12:03 — page 57 — #65 M. Y. Kizmaz 57 [5] Z. I. Borevich, Periodicity of residues of the number of finite labeled T0-topologies, (Russian) Modules and algebraic groups. Zap. Nauchn. Sem. Leningrad. Otdel. Mat. Inst. Steklov. (LOMI) 114 (1982), 32-36, 217. [6] M. Erné, Struktur- und Anzahlformeln für Topologien auf endlichen Mengen, Manuscr. Math.,11, No. 3, 221-259 (1974). [7] A. S. Mashhour, M. E. Abd El-Monsef and A. S. Farrag, On the number of topologies on a finite set, Delta J. Sci. 10 (1986), no. 1, 41-65. [8] R. Stanley, On the number of open sets of finite topologies, J. Combinatorial Theory 10 (1971) 74-79. Contact information M. Y. Kizmaz Department of Mathematics, Middle East Technical University, Ankara 06531, Turkey E-Mail(s): yasir@metu.edu.tr Received by the editors: 31.03.2017 and in final form 06.10.2017.
id nasplib_isofts_kiev_ua-123456789-188421
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
issn 1726-3255
language English
last_indexed 2025-12-07T16:09:39Z
publishDate 2019
publisher Інститут прикладної математики і механіки НАН України
record_format dspace
spelling Kizmaz, M.Y.
2023-02-28T18:51:11Z
2023-02-28T18:51:11Z
2019
On the number of topologies on a finite set / M.Y. Kizmaz // Algebra and Discrete Mathematics. — 2019. — Vol. 27, № 1. — С. 50–57. — Бібліогр.: 8 назв. — англ.
1726-3255
2010 MSC: Primary 11B50, Secondary 11B05.
https://nasplib.isofts.kiev.ua/handle/123456789/188421
We denote the number of distinct topologies which can be defined on a set X with n elements by T(n). Similarly, T0(n) denotes the number of distinct T₀ topologies on the set X. In the present paper, we prove that for any prime p, T(pᵏ) ≡ k + 1 (mod p), and that for each natural number n there exists a unique k such that T(p + n) ≡ k (mod p). We calculate k for n = 0, 1, 2, 3, 4. We give an alternative proof for a result of Z. I. Borevich to the effect that T₀(p + n) ≡ T₀(n + 1) (mod p).
en
Інститут прикладної математики і механіки НАН України
Algebra and Discrete Mathematics
On the number of topologies on a finite set
Article
published earlier
spellingShingle On the number of topologies on a finite set
Kizmaz, M.Y.
title On the number of topologies on a finite set
title_full On the number of topologies on a finite set
title_fullStr On the number of topologies on a finite set
title_full_unstemmed On the number of topologies on a finite set
title_short On the number of topologies on a finite set
title_sort on the number of topologies on a finite set
url https://nasplib.isofts.kiev.ua/handle/123456789/188421
work_keys_str_mv AT kizmazmy onthenumberoftopologiesonafiniteset