On stability of the Erdős-Rademacher Problem
Abstract.
Mantel’s theorem states that every -vertex graph with edges, where , contains a triangle. The problem of determining the minimum number of triangles in such a graph is usually referred to as the Erdős-Rademacher problem. Lovász and Simonovits proved that there are at least triangles in each of those graphs. Katona and Xiao considered the same problem under the additional condition that there are no vertices covering all triangles. They settled the case and . Solving their conjecture, we determine the minimum number of triangles for every fixed pair of and , when is sufficiently large. Additionally, solving another conjecture of Katona and Xiao, we extend the theory for considering cliques instead of triangles.
1991 Mathematics Subject Classification:
05C351. Introduction
A classical result in extremal combinatorics is Mantel’s theorem [11]. It says that every -vertex graph with at least edges contains a triangle. There have been various extensions of Mantel’s theorem. Turán [13] generalized it for cliques: Every -vertex graph with at least edges contains a clique of size , where is the number of edges of the complete balanced -partite graph on vertices. A result of Rademacher, see [3], says that every graph on vertices with edges contains at least triangles and this is best possible. Erdős [3, 4] conjectured that an -vertex graph with edges for contains at least triangles. This was proven by Lovász and Simonovits [10]. Xiao and Katona [14] proved a stability variant of the Lovász and Simonovits result for : If there is no vertex contained in all triangles, then there are at least triangles in . We will prove two extensions of this statement, one verifying a conjecture by Xiao and Katona [14] and the other proving a slight modification of another conjecture by Xiao and Katona [14].
For a graph denote the number of edges of and the number of copies of in . A -covering set in is a vertex set that contains at least one vertex of every copy of in . The -clique covering number is the size of the smallest -covering set. A triangle covering set is a -covering set and the triangle covering number is the -clique covering number. Given a vertex partition , we call an edge a class-edge if for some and a cross-edge otherwise. A set of two vertices is a missing cross-edge if and for some . Denote the number of cross-edges and the number of missing cross-edges. We drop the index and just write and , respectively, if is clear from context.
Conjecture 1 (Xiao, Katona [14]).
Let such that . Suppose the graph has vertices and edges satisfying and is large. Then contains at least
many triangles.
Xiao and Katona [14] proved their conjecture for and and that it is best possible. However, their conjecture in full generality holds only up to an additive constant. We will determine precisely the minimum number of triangles a graph on vertices with edges and can have. Depending on and one of the following two constructions will be an extremal example.
Construction 1: Let be the graph on vertices, where with
where is chosen later to minimize the number of triangles in this construction. Pick vertices in and two vertices in . Add the edges and to and delete the edges , where . Then, has edges, triangle covering number and
(1) |
Now, choose satisfying and minimizing (1). Note that is similar to the construction of Xiao and Katona [14] resulting in Conjecture 1, except that the class sizes are different in some cases.
Construction 2: Let be the graph on vertices, where with
where is chosen later to minimize the number of triangles in this construction. Pick vertices and one vertex . Add the edges to and delete the edges , where . The graph has edges, triangle covering number and
(2) |
Now, choose satisfying and minimizing (2). Our first main result is the following.
Theorem 1.
Let such that . Suppose the graph has vertices and edges, and is large. Then contains at least triangles.
Note that both and achieve this minimum for some values of and .
Xiao and Katona [14] also conjectured what the minimum number of copies of in a graph on vertices with edges and could be. They conjectured the following graph to be an extremal example.
Construction 3:
The graph is constructed as follows from the complete -partite graph with vertex partition where and . Let and . Remove the edge and add the edges and . The graph satisfies , and
We will verify Xiao and Katona’s [14] conjecture.
Theorem 2.
Let and be a sufficiently large integer. If a graph on vertices has edges and the copies of have an empty intersection, then the number of copies of is at least as many as in .
Theorem 3.
Let with and let be sufficiently large. Let be an -vertex graph with edges satisfying . Denote the family of graphs on vertices with edges such that there exists a vertex partition satisfying
-
•
the class-edges form a matching of size ,
-
•
,
-
•
all missing cross-edges are incident to one of the class-edges,
-
•
if the class-edges are not all in the same class, then each missing cross-edge is incident to class-edges on both endpoints.
Then for some .
We will observe (Lemma 4) that for all . Hence, the family is small. An optimization argument will reduce the family to and in the case , and to in the case . A similar cleaning argument could potentially be done in the general case, however the computational effort needed seems not worth the outcome. Also, it is likely that for some parameters , the extremal family realizing the minimum number of cliques of size contains several graphs.
For we know from a result of Erdős [5] that the number of cliques of size is at least and the graph constructed by adding a matching of size to one of the classes of the complete balanced -partite graph on vertices satisfies and . Hence this problem is interesting only when .
Our result can be extended to consider and as functions of . In particular, in the triangle case , our proof allows to be linear in . However, the methods we use, especially our main tool Theorem 6, will not give us the entire range of where our theorem should hold. Thus, we do not attempt to optimize our proof with respect to the dependency of and .
Our main tool is a stability-supersaturation result by Balogh, Bushaw, Collares, Liu, Morris and Sharifzadeh [1], which extends on the Erdős-Simonovits stability theorem [6] and Füredi’s [7] proof’s of it.
We would like to point out that Liu and Mubayi [9] independently obtained similar results to ours.
2. Proof of Theorem 3
The well-known Turán’s theorem [13] determines the maximum number of edges in a -free graph to be the number of edges of the complete balanced -partite graph. We will make use of the following bound on .
(3) |
The classical Erdős-Simonovits stability theorem [6] asserts that any -vertex -free graph with almost edges is close to the complete balanced -partite graph. There are many different quantitative versions [2, 8, 12] of this. A standard calculation shows that -vertex graphs with almost edges and a vertex partition with few class-edges need to have almost balanced class sizes.
Lemma 4.
Let , and a graph on vertices and edges with vertex partition such that the number of class-edges is . Then for all .
Proof.
Lemma 5.
Let , and a graph on vertices and edges with vertex partition such that the number of class-edges is . Then
Proof.
Any copy of needs to contain at least one class-edge. There are exactly copies of containing one given class-edge but no others, since the other vertices need to be chosen from different classes. Since the number of missing cross-edges is , they do not have an impact in this counting. There are class-edges and thus we conclude that the number of copies of containing exactly one class-edge is
The number of copies containing at least two class-edges is , because there are ways to chose three vertices among the vertices being incident to the class-edges and at most ways to chose the remaining vertices. We conclude
We say that a graph is -far from being -partite if for every subgraph with . Balogh, Bushaw, Collares, Liu, Morris and Sharifzadeh [1] proved that graphs which are -far from being -partite, contain many cliques of size .
Theorem 6 ([1]).
For every , the following holds. Every graph on vertices which is -far from being -partite contains at least
copies of .
With this tool in hand, we now prove Theorem 3.
Proof of Theorem 3.
Let be a graph on vertices and edges such that and is large enough. If is -far from being -partite, then by applying Theorem 6, the number of copies of in is at least
where we used (3) to lower bound . By Lemma 5, for every , and thus . Hence, we can assume that is not -far from being -partite. Then there exists a subgraph with and
Let be a partition such that are independent sets in . By Lemma 4, the class sizes are roughly balanced, that is for all .
Since , and , the vertex partition contains at most class-edges in . If the number of class-edges in is less than , then there is a -covering set of size at most , contradicting . If the number of class-edges is more than , then by Lemma 5
for every . Therefore, we can assume that the number of class-edges in is exactly . These edges need to form a matching, otherwise there is a -covering set of size , contradicting . Further, , because otherwise
Next, we assume that there is a missing cross-edge not incident to any of the class-edges. Take a cross-edge which is incident to exactly one class-edge. We will show that . Adding to increases the number of cliques of size by at most , because for each of the class-edges the number of cliques of size containing and is increased by at most . Removing from decreases the number of cliques of size by at least
where we used that in each class the number of vertices incident to no missing cross-edge is at least .
Thus, performing this edge flip decreases the number of cliques of size , i.e. . We repeat this process until we end up with a graph such that and every missing cross-edge is incident to a class-edge.
If has all class-edges in the same class, then . Thus, we can assume that has the property that not all class-edges are in the same class and there is a missing cross-edge not adjacent to class-edges on both ends. Denote the graph constructed from by adding all missing cross-edges. This graph satisfies
because for each class-edge and each added edge , the number of cliques of size containing and increases by at most if and share a vertex and by at most if and are disjoint.
There is a set of cross-edges such that each is incident to class-edges on both endpoints and every class-edge is incident to edges from on at most one endpoint. Denote the graph constructed from by removing those edges. This graph satisfies and
completing the proof. ∎
3. Proof of Theorem 1
Proof of Theorem 1.
By Theorem 3, we can assume that has a vertex partition where with exactly class-edges forming a matching. Let for some . We have
and thus the number of missing cross-edges is
Let be the matching of class-edges inside and be the matching of class-edges inside . Denote the graph constructed from by adding all missing cross-edges . The number of triangles in this graph is .
Case 1: .
The number of triangles in is at least , because each missing cross-edge is in at most two triangles in . Since , we get
Case 2: .
The number of triangles in is at least , because each missing cross-edge is in at most one triangle in . Therefore
4. Proof of Theorem 2
Let and be an integer such that . The number of cliques of size in is
Proof of Theorem 2.
Let be an -vertex graph with edges satisfying . By applying Theorem 3 for and , we can assume without loss of generality that and thus has a vertex partition where with exactly two class-edges and at most one missing cross-edge. We have
(5) |
This means we have two cases:
Case 1: .
In this case
by (5). Note that , because if , then moving one vertex from to increases the sum by at least two and thus
(6) |
a contradiction. If then , giving a contradiction as well. Thus, we have . Similarly, , as otherwise (6) holds. Conclusively, there are only the following types of combination of class sizes. We list them depending on . If , then
-
•
, , .
If , then
-
•
Type 1: , ,
-
•
Type 2: , , .
If , then
-
•
Type 1: , , ,
-
•
Type 2: , , .
If , then
-
•
Type 1: , , ,
-
•
Type 2: , .
If , then
-
•
, , .
Denote the two class-edges . Let such that and . The number of copies of containing but not is . Similarly, the number of copies of containing but not is . Thus, the total number of copies of in is at least
(7) |
We will now check that for any possible combination of class sizes the right hand side in (7) is at least . We distinguish cases depending on . If , then
If , then
If , then
If , then
If , then
Finally, if , then
Thus, in Case 1, .
Case 2: .
Observe that in this case the class sizes of and are the same. Further, the number of missing cross-edges is exactly one. Now, we distinguish two cases depending on where and are. Case a is when both class-edges are inside the same class. In Case b, and are in different classes.
Case 2a: for some
The missing edge is incident to or . The second endpoint of the missing edge is in for some . Then
where the sum of the factors are the same in the two products, hence the product is larger when the factors are ‘closer’ to each other.
Case 2b: for some .
Since the missing edge is incident to both class-edges,
where the last inequality holds by the same reasoning as in Case a.
∎
Acknowledgement
We would like to thank an anonymous referee for helpful comments, in particular, for pointing out an error in an earlier version of this paper.
References
- [1] J. Balogh, N. Bushaw, M. Collares, H. Liu, R. Morris, and M. Sharifzadeh. The typical structure of graphs with no large cliques. Combinatorica, 37(4):617–632, 2017.
- [2] J. Balogh, F. C. Clemen, M. Lavrov, B. Lidický, and F. Pfender. Making -free graphs -partite. Combinatorics, Probability and Computing, pages 1–10, 2020.
- [3] P. Erdős. Some theorems on graphs. Riveon Lematematika, 9:13–17, 1955.
- [4] P. Erdős. On a theorem of Rademacher-Turán. Illinois J. Math., 6:122–127, 1962.
- [5] P. Erdős. On the number of complete subgraphs and circuits contained in graphs. Časopis Pěst. Mat., 94:290–296, 1969.
- [6] P. Erdős and M. Simonovits. A limit theorem in graph theory. Studia Sci. Math. Hungar, 1:51–57, 1966.
- [7] Z. Füredi. A proof of the stability of extremal graphs, Simonovits’ stability from Szemerédi’s regularity. J. Combin. Theory Ser. B, 115:66–71, 2015.
- [8] D. Korándi, A. Roberts, and A. Scott. Exact stability for Turán’s theorem. arXiv:2004.10685, 2020.
- [9] X. Liu and D. Mubayi. On a generalized Erdős-Rademacher problem. arXiv:2005.07224, 2020.
- [10] L. Lovász and M. Simonovits. On the number of complete subgraphs of a graph. II. In Studies in pure mathematics, pages 459–495. Birkhäuser, Basel, 1983.
- [11] W. Mantel. Problem 28. Wiskundige Opgaven, 10:60–61, 1907.
- [12] A. Roberts and A. Scott. Stability results for graphs with a critical edge. arXiv:1610.08389, 2018.
- [13] P. Turán. Eine Extremalaufgabe aus der Graphentheorie. Mat. Fiz. Lapok, 48:436–452, 1941.
- [14] C. Xiao and G. O. H. Katona. The number of triangles is more when they have no common vertex. arXiv:2003.04450, 2020.