Smaller subgraphs of minimum degree k
Frank Mousset, Andreas Noever, and Nemanja \v{S}kori\'c

TL;DR
This paper advances the understanding of subgraphs with minimum degree k by showing that a linear number of vertices can be removed while maintaining the degree condition, moving closer to a longstanding conjecture.
Contribution
The authors improve existing bounds by demonstrating that at least a logarithmic fraction of vertices can be removed to preserve minimum degree k, supporting Erdős et al.'s conjecture.
Findings
At least rac{n}{\u221a{ ext{log} n}} vertices can be removed.
Progress towards Erdf3s et al.'s conjecture on subgraph size.
Enhanced bounds on subgraph vertex removal for minimum degree k.
Abstract
In 1990 Erd\H{o}s, Faudree, Rousseau and Schelp proved that for , every graph with vertices and edges contains a subgraph of minimum degree on at most vertices. They conjectured that it is possible to remove at least many vertices and remain with a subgraph of minimum degree , for some . We make progress towards their conjecture by showing that one can remove at least many vertices.
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.
Smaller subgraphs of minimum degree k
Frank Mousset
Andreas Noever
and Nemanja Škorić
Abstract
In 1990 Erdős, Faudree, Rousseau and Schelp proved that for , every graph with vertices and edges contains a subgraph of minimum degree on at most vertices. They conjectured that it is possible to remove at least many vertices and remain with a subgraph of minimum degree , for some . We make progress towards their conjecture by showing that one can remove at least many vertices.
1 Introduction
It is easy to show that every graph on vertices with at least edges contains a subgraph of minimum degree . More generally, any graph on vertices with at least edges111Technically, we may relax the first condition to , since for the condition on the number of edges cannot be satisfied. However, there seems no point in doing so. For , the statement is wrong. contains a subgraph of minimum degree , for all . This statement is best possible in two ways: (1) there exist graphs on vertices with edges which do not contain a subgraph of minimum degree , and (2) there exist graphs on vertices with edges without a subgraph of minimum degree on fewer than vertices. For example the wheel (where denotes the graph join operation) has exactly edges and minimum degree , but contains no proper induced subgraph with minimum degree . A similar construction is available for all (consider the generalized wheel ).
Erdős conjectured that the presence of even a single additional edge allows one to find a much smaller subgraph:
Conjecture 1.1** (Erdős [1, 2]).**
For every there exists an such that every graph on vertices and edges contains a subgraph of minimum degree with at most vertices.
As far as we know, the only progress on this conjecture is the following theorem due to Erdős, Faudree, Rousseau and Schelp from 1990:
Theorem 1.2** (Erdős, Faudree, Rousseau, Schelp [2]).**
For , let be a graph on vertices and edges. Then contains a subgraph of order at most and minimum degree at least .
Here, we will show that it is possible to replace the by :
Theorem 1.3**.**
For , let be a graph on vertices and edges. Then contains a subgraph of order at most and minimum degree at least .
We use standard graph theoretic notation. The vertex and edge sets of a graph are denoted by and . We write for the number of vertices and for the number of edges in . denotes the set of vertices of with degree exactly . Similarly denotes the set of vertices of degree at most . For a vertex , we denote its neighborhood by and its degree by (we omit the subscript if it is clear from the context). We write (or ) for the set of edges in with one endpoint in and another in . The minimum degree of a graph is denoted by .
2 Proof of Theorem 1.3
The proof will use induction on the number of vertices of . In the base case it is easy to check that , so the theorem holds vacuously. Assume now that the theorem holds for all graphs on vertices.
If contains a vertex with , then has at least edges. Thus, by induction, (and hence ) contains a subgraph with minimum degree and at most
[TABLE]
vertices, and we are done. From now on, we will assume that has minimum degree at least .
The rest of the proof is split into two cases depending on the number of vertices of degree exactly . Already in the proof of Theorem 1.2, Erdős, Faudree, Rousseau, and Schelp observed that Conjecture 1.1 holds if the number of vertices of degree is not too large:
Lemma 2.1** (Lemma 4 in [2]).**
For , let be a graph on vertices and edges with . If for some , has at most vertices of degree , then has a subgraph of order at most with .
Set . The exact value of is not too important, but we need and the given value seems convenient. If contains fewer than vertices of degree then, by 2.1, contains a subgraph of order at most
[TABLE]
with minimum degree , and we are done. So from now on assume that contains at least vertices of degree . The following notion is important for the rest of the proof:
Definition 2.2** (Good set).**
By a good set, we mean any set of vertices of constructed according to the following rules:
If has degree , then is good. 2. 2.
If is good and is such that all but at most neighbors of belong to , then is good. 3. 3.
If and are both good and if contains an edge that meets both and , then is good (this is the case if or if ).
We say that a good set is maximal if it is not properly contained in another good set.
The relevance of this notion for our problem is partly due to the following claim.
Claim 2.3**.**
The following statements hold:
- (i)
Every good set intersects at most edges of . 2. (ii)
If is a good set with , then contains a subgraph of minimum degree at least . 3. (iii)
If and are maximal good sets and , then and .
Proof.
The statement (i) is easily proved by induction on the rules 1, 2, and 3 in the definition of a good set. For (ii), suppose that , so . Because intersects at most edges, we have
[TABLE]
where the last equality follows from the definition . Then (ii) follows because every graph with and contains a subgraph of minimum of minimum degree . Statement (iii) follows immediately from the definition of a good set (rule 3). ∎
We now handle the case where some good set is very large. Since every good set is obtained by application of one of the rules given in Definition 2.2, it is clear that if is a good set of size at least two, then there is a good subset of size . In particular, if some good set has size at least , then there also exists a good set satisfying , where the last inequality holds since . Then Claim 2.3 (ii) implies that has a subgraph of minimum degree , and so has a subgraph of order at most with minimum degree , and we are done. From now on, we may thus assume that every good set has size at most .
To motivate the rest of the proof, we now briefly discuss the proof strategy that was used by Erdős, Faudree, Rousseau, and Schelp in [2]. Let be some parameter with , which we will optimize later. If there is a good set of size at least , then Claim 2.3 (ii) implies that we can find a subgraph of minimum degree on at most vertices (the case where the good set is larger than is already excluded). Otherwise, every good set is smaller than . Now we run an algorithm which constructs a chain of subgraphs greedily as follows: if contains a maximal good set such that has minimum degree at least , then we let ; otherwise . One can show, using the statements in Claim 2.3 and the fact that every maximal good set has size at most , that this algorithm removes at least good sets, and thus produces a graph with vertices, which has minimum degree at least by construction. By setting to the optimal value , we thus obtain a subgraph with minimum degree of size . In our proof, we avoid the case distinction based on the maximum size of a good set. However, we also construct a small subgraph of by removing maximal good sets. More precisely, we start by choosing a collection of maximal good sets that covers vertices and such that all good sets in are of comparable sizes; the existence of this collection is guaranteed by Claim 2.4 further below. Next, we remove a positive fraction of the sets in from in such a way that the remaining graph still contains a subgraph of minimum degree . The main technical statement that makes this possible is Claim 2.5 below. Since the sets in all have similar sizes, this means that we remove vertices, completing the proof. We now turn to the details.
Claim 2.4**.**
There exists a collection of maximal good sets such that
[TABLE]
and such that for any two , we have .
Proof.
Let denote a maximum-size collection of maximal good sets. Since contains at least vertices of degree , and since by maximality of , every such vertex is contained in one of the good sets in , we have . For every , let denote the subfamily of all with . By the pigeonhole principle, there exists such that . If then is a collection with the desired properties. Otherwise we remove a single good set, say , from . Since each good set has size at most , we have
[TABLE]
where we used . Then is a collection with the desired properties. ∎
Let be a collection as in the statement of Claim 2.4. For every , we define to be the set of all good sets such that contains a neighbor of in . Moreover, we let be any subcollection of size if and we let otherwise.
Claim 2.5**.**
There is a set of size at most with the following property: for every subfamily such that for all , we have , the graph contains a subgraph of minimum degree .
We postpone the proof of Claim 2.5 to the end of the proof. For now, assume that we have a set as in the claim. Our goal is to find a subcollection of size containing at most one good set from every . To find such a collection , we construct an auxiliary ‘conflict graph’ on the vertex set by adding a clique on for every . Note that we are looking for an independent set in . Because holds by construction and since , we have
[TABLE]
By Turán’s theorem, any graph on vertices with at most edges contains an independent set of size at least . Thus, contains an independent set of size at least . Because any two satisfy and since , we have
[TABLE]
Then has at most vertices and contains a subgraph of minimum degree , by the defining property of . Recalling that , this completes the proof of the theorem. It remains to prove Claim 2.5.
2.1 Proof of Claim 2.5
For the proof of the claim, we need the following definition and lemma.
Definition 2.6** (-cover).**
Suppose that is a graph and that is a subset of its vertices. Given , a graph is called an -cover if it contains as a subgraph and if .
Lemma 2.7**.**
Suppose that is a graph that does not contain a subgraph of minimum degree , for . Then there exists a subset of cardinality at most such that every -cover contains a subgraph of minimum degree .
Proof.
For a graph , we define the function
[TABLE]
By induction on the number of vertices of , we will prove the following statement, which is slightly stronger than the claim of the lemma: if has no subgraph of minimum degree , then there is a subset of size at most such that every -cover contains a subgraph of minimum degree .
In the base case , we can let be the set containing the single vertex of . In this case, we have , and every -cover has minimum degree at least by definition.
If then the fact that does not contain a subgraph of minimum degree implies that there is a vertex with . Let us write . Then is a graph without a subgraph of minimum degree , and it has fewer vertices than . Hence, by induction, it contains a set such that every -cover has a subgraph of minimum degree . We now define the set by
[TABLE]
where
[TABLE]
Note that from the definition of it follows that if then we have . Furthermore, since and by definition does not contain vertices of degree , we also have . To check that is not too big, note that
[TABLE]
Therefore if then . If then we distinguish two cases:
If and then there must exist a vertex with degree at least . By definition of , has degree at most in and thus has degree exactly in . In particular, we have and thus . 2. 2.
Otherwise, we either have or we have and . In both cases, we have and so .
Now suppose that is an -cover. We claim that then either is an -cover or is an -cover. In both cases, the induction hypothesis implies that has a subgraph of minimum degree . To show this, we first recall that by the definition of an -cover, we have . We distinguish two cases. If then actually . Moreover, by construction of , all vertices of that belong to must have degree in (and so degree at least in ). Therefore we have . In other words, is -cover. Otherwise, we have . Then and thus , and . It moreover follows that . These observations show that
[TABLE]
Thus is an -cover, completing the proof. ∎
Using this lemma, we now prove Claim 2.5. Consider the graph obtained by removing all sets in from . We have . By Claim 2.3 (i), every good set intersects at most edges, and so
[TABLE]
If contains a subgraph of minimum degree , then we are done because we can simply choose . Otherwise, we apply 2.7 to to obtain a set of size
[TABLE]
(using (2) and the definition of to bound ) such that every -cover contains a subgraph of minimum degree . To complete the proof of the claim, suppose that contains at most one set from each set where . It is enough to show that the graph is an -cover. Note first that since each element of is a maximal good set, the elements of are pairwise disjoint and for any two distinct , there are no edges between and in . Since has minimum degree at least by assumption, this means in particular that . Furthermore, every has degree at least in , for one of the following two reasons:
Either or in which case the claim directly follows. 2. 2.
Or and we have removed exactly one good set from , say . Then the degree of in (which equals the degree in ) must be at least by the maximality of .
Thus , so is an -cover. This completes the proof of the claim.
3 Acknowledgements
We are grateful to Rajko Nenadov and Yury Person for fruitful discussions.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] P. Erdős. Some of my favorite solved and unsolved problems in graph theory. Quaestiones Mathematicae , 16(3):333--350, 1993.
- 2[2] P. Erdős, R. Faudree, C. Rousseau, and R. Schelp. Subgraphs of minimal degree k. Discrete Mathematics , 85(1):53 -- 58, 1990.
