A degree version of the Hilton--Milner theorem
Peter Frankl, Jie Han, Hao Huang, Yi Zhao

TL;DR
This paper extends the Hilton--Milner theorem by establishing a degree version for non-trivial intersecting families of sets, showing that the minimum degree is maximized by a specific extremal family.
Contribution
It introduces a degree-based stability result for intersecting families, generalizing the classical Hilton--Milner theorem with a novel proof approach.
Findings
Minimum degree of non-trivial intersecting families is bounded by that of a specific extremal family.
The degree version holds when n=Ω(k^2), extending previous size-based results.
Uses fundamental extremal set theory results and introduces a new variant of the Erd ext{"o}s--Ko--Rado theorem.
Abstract
An intersecting family of sets is trivial if all of its members share a common element. Hilton and Milner proved a strong stability result for the celebrated Erd\H{o}s--Ko--Rado theorem: when , every non-trivial intersecting family of -subsets of has at most members. One extremal family consists of a -set and all -subsets of containing a fixed element and at least one element of . We prove a degree version of the Hilton--Milner theorem: if and is a non-trivial intersecting family of -subsets of , then , where denotes the minimum (vertex) degree of . Our proof uses several fundamental results in extremal set theory, the concept of kernels, and a new variant…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsLimits and Structures in Graph Theory · Advanced Topology and Set Theory · Functional Equations Stability Results
A degree version of the Hilton–Milner theorem
Peter Frankl
Alfréd Rényi Institute of Mathematics, P.O.Box 127, H-1364 Budapest, Hungary
,
Jie Han
Instituto de Matemática e Estatística, Universidade de São Paulo, Rua do Matão 1010, 05508-090, São Paulo, Brazil
,
Hao Huang
Department of Math and CS, Emory University, Atlanta, GA 30322
and
Yi Zhao
Department of Mathematics and Statistics, Georgia State University, Atlanta, GA 30303
Abstract.
An intersecting family of sets is trivial if all of its members share a common element. Hilton and Milner proved a strong stability result for the celebrated Erdős–Ko–Rado theorem: when , every non-trivial intersecting family of -subsets of has at most members. One extremal family consists of a -set and all -subsets of containing a fixed element and at least one element of . We prove a degree version of the Hilton–Milner theorem: if and is a non-trivial intersecting family of -subsets of , then , where denotes the minimum (vertex) degree of . Our proof uses several fundamental results in extremal set theory, the concept of kernels, and a new variant of the Erdős–Ko–Rado theorem.
Key words and phrases:
Intersecting families, Hilton–Milner theorem, Erdős–Ko–Rado theorem
2010 Mathematics Subject Classification:
05D05
JH is supported by FAPESP (Proc. 2013/03447-6, 2014/18641-5). HH is partially supported by a Simons Collaboration Grant. YZ is partially supported by NSF grant DMS-1400073.
1. Introduction
A family of sets is called intersecting if for all . A fundamental problem in extremal set theory is to study the properties of intersecting families. For positive integers , let and denote the family of all -element subsets (-subsets) of . We call a family on -uniform if it is a subfamily of . A full star is a family that consists of all the -subsets of that contains a fixed element. We call an intersecting family trivial if it is a subfamily of a full star. The celebrated Erdős–Ko–Rado (EKR) theorem [3] states that, when , every -uniform intersecting family on has at most members, and the full star shows that the bound is best possible. Hilton and Milner [14] proved the uniqueness of the extremal family in a stronger sense: if , every non-trivial intersecting family of -subsets of has at most members. It is easy to see that the equality holds for the following family, denoted by , which consists of a -set and all -subsets of containing a fixed element and at least one vertex of . For more results on intersecting families, see a recent survey by Frankl and Tokushige [10].
Given a family and , we denote by the subfamily of consisting of all the members of that contain , i.e., . Let be the degree of . Let and denote the maximum and minimum degree of , respectively. There were extremal problems in set theory that considered the maximum or minimum degree of families satisfying certain properties. For example, Frankl [7] extended the Hilton–Milner theorem by giving sharp upper bounds on the size of intersecting families with certain maximum degree. Bollobás, Daykin, and Erdős [1] studied the minimum degree version of a well-known conjecture of Erdős [2] on matchings.
Huang and Zhao [15] recently proved a minimum degree version of the EKR theorem, which states that, if and is a -uniform intersecting family on , then , and the equality holds only if is a full star. This result implies the EKR theorem immediately: given a -uniform intersecting family , by recursively deleting elements with the smallest degree until elements are left, we derive that
[TABLE]
Frankl and Tokushige [11] gave a different proof of the result of [15] for . Generally speaking, a minimum degree condition forces the sets of a family to be distributed somewhat evenly and thus the size of a family that is required to satisfy a property might be smaller than the one without degree condition. Unless the extremal family is very regular, an extremal problem under the minimum degree condition seems harder than the original extremal problem because one cannot directly apply the shifting method (a powerful tool in extremal set theory).
In this paper we study the minimum degree version of the Hilton–Milner theorem.
Theorem 1**.**
Suppose and , where for and for . If is a non-trivial intersecting family, then .
Han and Kohayakawa [12] recently determined the maximum size of a non-trivial intersecting family that is not a subfamily of , which is . Later Kostochka and Mubayi [17] determined the maximum size of a non-trivial intersecting family that is not a subfamily of or the extremal families given in [12] for sufficiently large . Furthermore, Kostochka and Mubayi [17, Theorem 8] characterized all maximal intersecting -uniform families on for and . Using a different approach, Polcyn and Ruciński [18, Theorem 4] characterized all maximal intersecting -uniform families on for , in particular, there are fifteen such families, including the full star and . It is straightforward to check that all these families have minimum degree at most – this gives the following proposition.
Proposition 2**.**
If and is a non-trivial intersecting family, then .
In order to prove Theorem 1, we prove a new variant of the EKR theorem, which is closely related to the EKR theorem for direct products given by Frankl (see Theorem 7).
Theorem 3**.**
Given integers , , and , let be three disjoint -subsets of . If is a -uniform intersecting family on such that every member intersects all of , then .
Theorem 3 becomes trivial when because every family of -sets that intersect satisfies . Our bound in Theorem 3 is asymptotically tight because a star with a center in contains about -sets that intersect .
It was shown in [15] that one can derive the minimum degree version of the EKR theorem for by using the Hilton–Milner Theorem and simple averaging arguments (thus the difficulty of the result in [15] lies in deriving the tight bound ). However, we can not use this naive approach to prove Theorem 1 for sufficiently large . Indeed, let be a non-trivial intersecting family that is not a subfamily of . The result of Han and Kohayakawa [12] says that is asymptotically at most , and in turn, the average degree of is asymptotically at most . Unfortunately, this is much larger than as is fixed and is sufficiently large.
Our proof of Theorem 1 applies several fundamental results in extremal set theory as well as Theorem 3. The following is an outline of our proof. Let be a non-trivial intersecting family such that . For every , we obtain a lower bound for by applying the assumption on and the Frankl–Wilson theorem [5, 19] on the maximum size of -intersecting families. If , then we derive a contradiction by considering the kernel of (a concept introduced by Frankl [6]). When , we separate two cases based on . When is large, assume that and let . A result of Frankl [9] implies that contains three edges , , where are pairwise disjoint. Since is intersecting and every member of meets each of , Theorem 3 gives an upper bound on , which contradicts the lower bound that we derived earlier. When is small, we apply the aforementioned result of Frankl [7] to obtain an upper bound on , which contradicts the assumption on .
2. Tools
2.1. Results that we need
Given a positive integer , a family of sets is called -intersecting if for all . A -intersecting EKR theorem was proved in [3] for sufficiently large . Later Frankl [5] (for ) and Wilson [19] (for all ) determined the exact threshold for .
Theorem 4**.**
[5, 19]** Let and let be a -uniform -intersecting families on . Then .
As mentioned in Section 1, Frankl [7] determined the maximum possible size of an intersecting family under a maximum degree condition.
Theorem 5**.**
[7]** Suppose , , is intersecting. If , then .
Given a -uniform family , a matching of size is a collection of vertex-disjoint sets of . A well-known conjecture of Erdős [2] states that if and satisfies , then contains a matching of size . Frankl [9] verified this conjecture for .
Theorem 6**.**
[9]** Let and let . If , then contains a matching of size .
Frankl [8] proved an EKR theorem for direct products.
Theorem 7**.**
[8]** Suppose and , where are positive integers. Let be a partition of with , and
[TABLE]
If for all and is intersecting, then
[TABLE]
Note that the case of Theorem 7 is the EKR theorem.
2.2. Kernels of intersecting families
Frankl introduced the concept of kernels (and called them bases) for intersecting families in [6]. Given , a set is called a cover of if for all . For example, if is intersecting, then every member of is a cover. Given an intersecting family , we define its kernel as
[TABLE]
An intersecting family is called maximal if is not intersecting for any -set . Note that, when proving Theorem 1, we may assume that is maximal because otherwise we can add more -sets to such that the resulting intersecting family is still non-trivial and satisfies the minimum degree condition. We observe the following fact on the kernels.
Fact 8**.**
If and is a maximal intersecting family, then is also intersecting.
Proof.
Suppose there are such that . Since , we can find two disjoint -sets on such that for . For , since is a cover of , intersects all members of . Since is maximal, we derive that . This contradicts the assumption that are disjoint. ∎
For , let . If an intersecting family is non-trivial, then . Below we prove an upper bound for , , where the case was given by Erdős and Lovász [4].
Lemma 9**.**
For , we have .
In order to prove Lemma 9, We use a result of Håstad, Jukna, and Pudlák [13, Lemma 3.4]. Given a family , the cover number of , denoted by , is the size of the smallest cover of .
Lemma 10**.**
[13]** If is an -uniform family with , then there exists a set such that , where .
Proof of Lemma 9.
Suppose for some . Then by Lemma 10, there exists a set such that . In particular, is nonempty, namely, there exists such that . By the definition of , this implies that is not a cover of , so there exists such that . Since each member of is a cover of , each of them intersects . This implies that , a contradiction. ∎
3. Proof of Theorem 3
In this section we derive Theorem 3 from Theorem 7.
Proof of Theorem 3.
Let consist of all the subsets of that intersect with in exactly elements. Then . Let , , , , and , . Since , we have . Since , we can apply Theorem 7 to conclude that
[TABLE]
Note that a set intersects with either or or elements. We partition into three subfamilies accordingly. Our assumption implies
[TABLE]
We can apply Theorem 7 to each subfamily of and obtain that
[TABLE]
Finally, for , we claim that . Indeed, let , , and . Note that and . If , then Theorem 7 gives that
[TABLE]
When , we have because . Hence,
[TABLE]
and the trivial bound on gives that
[TABLE]
as claimed. Summing up the bounds for and for , we have
[TABLE]
because . ∎
4. Proof of Theorem 1
We start with some simple estimates. First, for , and , we have
[TABLE]
Similarly, one can show that . Second, if , then we have
[TABLE]
Lemma 11**.**
Suppose and , is a non-trivial intersecting family such that . Then for any ,
- ()
there exists such that and ; 2. ()
.
Proof.
Given , write and . If , then , and thus , a contradiction. So assume that .
Let among all distinct . Obviously , and is a -intersecting family on . Then since , we get by Theorem 4. Note that there exist such that . Since every set in must intersect both and , for every , by the inclusion-exclusion principle, we have
[TABLE]
Let and thus . Suppose attains the minimum degree in among all elements of . Since , by (4.3) we have
[TABLE]
By the definition of we get
[TABLE]
where the factor comes from the fact that every member is counted at most times – because . By (4.1) with and , we get
[TABLE]
which, together with , implies that , so () holds. Since , the first inequality above gives (). ∎
Proof of Theorem 1.
First assume that and . Suppose is a non-trivial intersecting family such that . Suppose attains the maximum degree of and write . If , then by Theorem 6, the -uniform family contains a matching of size . Every member of must intersect each of . By Theorem 3, we have . On the other hand, Lemma 11 Part () implies that because . This gives a contradiction.
We thus assume that . By Theorem 5,
[TABLE]
Since , by (4.2), we have . The upper and lower bounds for together imply , a contradiction.
Now assume that and . Since is intersecting, each member of is a cover of and thus contains as a subset a minimal cover, which is a member of the kernel . Thus . We know because is non-trivial. We observe that – otherwise assume (recall that is intersecting). By the definition of , every contains both and so every satisfy that , contradicting Lemma 11 Part (). By Lemma 9,
[TABLE]
Since , for any , we have
[TABLE]
Thus
[TABLE]
On the other hand, by (4.2), we have . Hence, , contradicting . This completes the proof of Theorem 1. ∎
5. Concluding Remarks
The main question arising from our work is whether Theorem 1 holds for all . Proposition 2 confirms this for . Another question is whether the following generalization of Theorems 3 and 7 is true. We say a family of sets has the EKR property if the largest intersecting subfamily of is trivial.
Conjecture 12**.**
Suppose and , where are integers. Let be a partition of with , and
[TABLE]
*If for all and for all but at most one such that , then has the EKR property. *
The assumptions on cannot be relaxed for the following reasons. If for some , then itself is intersecting and for any . If for distinct such that , then for any , the union of and is a larger intersecting family than .
When , Conjecture 12 follows from Theorem 7, in particular, the case is the EKR theorem. A recent result of Katona [16] confirms Conjecture 12 for the case and . We can prove Conjecture 12 in the following case.
Theorem 13**.**
Given positive integers , with , there exists such that the followings holds for all . If are disjoint subsets of such that for all , then
[TABLE]
*has the EKR property. *
We omit the proof of Theorem 13 here because the purpose of this paper is to prove Theorem 1. Moreover, when and , our is so we cannot replace Theorem 3 by Theorem 13 in our main proof. Nevertheless, it would be interesting to know the smallest such that Theorem 13 holds.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] B. Bollobás, D. E. Daykin, and P. Erdős. Sets of independent edges of a hypergraph. Quart. J. Math. Oxford Ser. (2) , 27(105):25–32, 1976.
- 2[2] P. Erdős. A problem on independent r 𝑟 r -tuples. Ann. Univ. Sci. Budapest. Eötvös Sect. Math. , 8:93–95, 1965.
- 3[3] P. Erdős, C. Ko, and R. Rado. Intersection theorems for systems of finite sets. Quart. J. Math. Oxford Ser. (2) , 12:313–320, 1961.
- 4[4] P. Erdős and L. Lovász. Problems and results on 3 3 3 -chromatic hypergraphs and some related questions. pages 609–627. Colloq. Math. Soc. János Bolyai, Vol. 10, 1975.
- 5[5] P. Frankl. The Erdős-Ko-Rado theorem is true for n = c k t 𝑛 𝑐 𝑘 𝑡 n=ckt . In Combinatorics (Proc. Fifth Hungarian Colloq., Keszthely, 1976), Vol. I , volume 18 of Colloq. Math. Soc. János Bolyai , pages 365–375. North-Holland, Amsterdam-New York, 1978.
- 6[6] P. Frankl. On intersecting families of finite sets. J. Combin. Theory Ser. A , 24(2):146 – 161, 1978.
- 7[7] P. Frankl. Erdős-Ko-Rado theorem with conditions on the maximal degree. J. Combin. Theory Ser. A , 46(2):252–263, 1987.
- 8[8] P. Frankl. An Erdő-Ko- Rado theorem for direct products. European J. Combin. , 17(8):727 – 730, 1996.
