On the super domination number of graphs
Douglas J. Klein, Juan A. Rodr\'iguez-Vel\'azquez, Eunjeong Yi

TL;DR
This paper investigates the super domination number in graphs, providing formulas and bounds, and explores specific cases like corona and Cartesian product graphs to deepen understanding of this graph invariant.
Contribution
It introduces closed formulas and tight bounds for the super domination number, including special cases for product graphs, advancing theoretical understanding of this graph parameter.
Findings
Derived closed formulas for super domination number.
Established tight bounds based on graph invariants.
Analyzed super domination in corona and Cartesian product graphs.
Abstract
The open neighbourhood of a vertex of a graph is the set consisting of all vertices adjacent to in . For , we define . A set is called a super dominating set of if for every vertex , there exists such that . The super domination number of is the minimum cardinality among all super dominating sets in . In this article, we obtain closed formulas and tight bounds for the super domination number of in terms of several invariants of . Furthermore, the particular cases of corona product graphs and Cartesian product graphs are considered.
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
TopicsAdvanced Graph Theory Research · Graph Labeling and Dimension Problems · Optimization and Search Problems
On the super domination number of graphs
Douglas J. Klein1, Juan A. Rodríguez-Velázquez1,2, Eunjeong Yi1
1Texas AM University at Galveston
Foundational Sciences, Galveston, TX 77553, United States
2Universitat Rovira i Virgili, Departament d’Enginyeria Informàtica i Matemàtiques
Av. Països Catalans 26, 43007 Tarragona, Spain
Email addresses: [email protected], [email protected],[email protected]
Abstract
The open neighbourhood of a vertex of a graph is the set consisting of all vertices adjacent to in . For , we define . A set is called a super dominating set of if for every vertex , there exists such that . The super domination number of is the minimum cardinality among all super dominating sets in . In this article, we obtain closed formulas and tight bounds for the super domination number of in terms of several invariants of . Furthermore, the particular cases of corona product graphs and Cartesian product graphs are considered.
Keywords: Super domination number; Domination number; Cartesian product; Corona product.
AMS Subject Classification numbers: 05C69; 05C70 ; 05C76
1 Introduction
The open neighbourhood of a vertex of a graph is the set consisting of all vertices adjacent to in . For , we define . A set is dominating in if every vertex in has at least one neighbour in , i.e., for every . The domination number of , denoted by , is the minimum cardinality among all dominating sets in . A dominating set of cardinality is called a -set. The reader is referred to the books [9, 10] for details on domination in graphs.
A set is called a super dominating set of if for every vertex , there exists such that
[TABLE]
If and satisfy (1), then we say that is a private neighbour of with respect to . The super domination number of , denoted by , is the minimum cardinality among all super dominating sets in . A super dominating set of cardinality is called a -set. The study of super domination in graphs was introduced in [14]. We recall some results on the extremal values of .
Theorem 1**.**
[14]*
Let be a graph of order . The following assertions hold.*
- •
* if and only if is an empty graph.*
- •
.
- •
* if and only if or .*
It is well known that for any graph without isolated vertices, . As noticed in [14], from the theorem above we have that for any graph without isolated vertices,
[TABLE]
Connected graphs with were characterized in [14], while all graphs with were characterized in [4].
The following examples were previously shown in [14].
- (a)
For a complete graph with , . 2. (b)
For a star graph . 3. (c)
For a complete bipartite graph with , .
The three cases above can be generalized as follows. Let be the complete -partite graph of order . If at most one value is greater than one, then as in such a case or , where denotes the join of graphs and . As shown in [4], these cases are included in the family of graphs with super domination number equal to . On the other hand, it is not difficult to show that if there are at least two values , then . We leave the details to the reader. In summary, we can state the following.
[TABLE]
The particular case of paths and cycles was studied in [14].
Theorem 2**.**
[14]* For any integer ,*
[TABLE]
[TABLE]
It was shown in [4] that the problem of computing is NP-hard. This suggests that finding the super domination number for special classes of graphs or obtaining good bounds on this invariant is worthy of investigation. In particular, the super domination number of lexicographic product graphs and joint graphs was studied in [4] and the case of rooted product graphs was studied in [13]. In this article we study the problem of finding exact values or sharp bounds for the super domination number of graphs.
The article is organised as follows. In Section 2, we study the relationships between and several parameters of , including the number of twin equivalence classes, the domination number, the secure domination number, the matching number, the -packing number, the vertex cover number, etc. In Section 3, we obtain a closed formula for the super domination number of any corona graph, while in Section 4 we study the problem of finding the exact values or sharp bounds for the super domination number of Cartesian product graphs and express these in terms of invariants of the factor graphs.
2 Relationship between the super domination number and other parameters of graphs
A matching, also called an independent edge set, on a graph is a set of edges of such that no two edges share a vertex in common. A largest matching (commonly known as a maximum matching or maximum independent edge set) exists for every graph. The size of this maximum matching is called the matching number and is denoted by .
Theorem 3**.**
For any graph of order ,
[TABLE]
Proof.
Let be a -set and let be a set of cardinality such that for every there exists such that Since
[TABLE]
is a matching, we have that as required. ∎
As a simple example of a graph where the bound above is achieved we can take . In this case , and Another example is whenever
A vertex cover of is a set such that each edge of is incident to at least one vertex of . A minimum vertex cover is a vertex cover of smallest possible cardinality. The vertex cover number is the cardinality of a minimum vertex cover of .
Theorem 4**.**
(König [12], Egerváry [5])* For bipartite graphs the size of a maximum matching equals the size of a minimum vertex cover.*
We use Theorems 3 and 4 to derive the following result.
Theorem 5**.**
For any bipartite graph of order ,
[TABLE]
The bound above is attained, for instance, for any star graph and for any hypercube graph. It is well-known that for the hypercube , (see, for instance, [8]) and in Section 4 we will show that .
An independent set of is a set such that no two vertices in are adjacent in , and the independence number of , , is the cardinality of a largest independent set of .
The following well-known result, due to Gallai, states the relationship between the independence number and the vertex cover number of a graph.
Theorem 6**.**
(Gallai’s theorem)* For any graph of order ,*
[TABLE]
From Theorems 5 and 6 we deduce the following tight bound.
Theorem 7**.**
For any bipartite graph of order ,
[TABLE]
A set is said to be a secure dominating set of if it is a dominating set and for every there exists such that is a dominating set. The secure domination number, denoted by , is the minimum cardinality among all secure dominating sets. This type of domination was introduced by Cockayne et al. in [3].
Theorem 8**.**
For any graph ,
[TABLE]
Proof.
Let be a -set. For each let such that . Obviously, and are dominating sets. Therefore, is a secure dominating set and so . ∎
The inequality above is tight. For instance, Moreover, the equality is also achieved whenever , since . This case will be discussed in Corollary 13.
The closed neighbourhood of a vertex is defined to be . We define the twin equivalence relation on as follows:
[TABLE]
We have three possibilities for each twin equivalence class :
- (a)
is a singleton set, or 2. (b)
and for any , or 3. (c)
and for any .
We will refer to the types (a) (b) and (c) classes as the singleton, false and true twin equivalence classes, respectively.
Let us see three different examples of non-singleton equivalence classes. An example of a graph where every equivalence class is a true twin equivalence class is , . In this case, there are three equivalence classes composed of and true twin vertices, respectively. As an example where every class is composed of false twin vertices, we take the complete bipartite graph , . Finally, the graph , , has two equivalence classes and one of them is composed of true twin vertices and the other one is composed of false twin vertices. On the other hand, , , is an example where one class is singleton, one class is composed of true twin vertices and the other one is composed of false twin vertices.
The following straightforward lemma will be very useful to prove our next theorem.
Lemma 9**.**
Let be a graph and a -set. Let such that and for every there exists such that If is a twin equivalence class, then and .
Theorem 10**.**
For any graph of order having twin equivalence classes,
[TABLE]
Furthermore, if is connected and , then
[TABLE]
Proof.
Let be the set of twin equivalence classes of and let be a -set. By Lemma 9 for every twin equivalence class we have , which implies that .
Suppose that . In such a case, by Lemma 9 we have that and for every twin equivalence class. From now on we assume that is connected and . Hence, there exist three twin equivalence classes such that every vertex in is adjacent to every vertex in , and six vertices , and such that and . Thus, , which is a contradiction. Therefore, the result follows. ∎
The bound is achieved by complete nontrivial graphs , complete bipartite graphs , where , and by the disjoin union of these graphs. The bound is achieved by the graph shown in Figure 1, where there are four false twin equivalence classes and a singleton equivalence class. In this case, white-coloured vertices form a -set.
The open neighbourhood of a set is defined to be , while the closed neighbourhood is defined to be . A set is open irredundant if for every ,
[TABLE]
Theorem 11**.**
[1]* If a graph has no isolated vertices, then has a minimum dominating set which is open irredundant.*
Theorem 12**.**
If a graph has no isolated vertices, then
[TABLE]
Proof.
If has no isolated vertices, by Theorem 11, there exists an open irredundant set such that . Since every satisfies (4), we have that for every , there exits such that , which implies that is a super dominating set. Therefore, ∎
The bound above is tight. For instance, it is achieved by the graph shown in Figure 2. In Section 3 we will show other examples of graphs where the bound above is achieved.
Since , we deduce the following consequence of Theorem 12.
Corollary 13**.**
Let be a graph of order . If , then .
The converse of Corollary 13 is not true. For instance, as we will see in Section 4, for the Cartesian product of times we have , while .
A set is called a -packing if for every pair of different vertices . The -packing number is the cardinality of any largest -packing of . It is well known that for any graph , . Meir and Moon [15] showed in 1975 that for any tree . We remark that in general, these -sets and -sets are not identical.
Corollary 14**.**
Let be a graph of order . If does not have isolated vertices,
[TABLE]
To show that the bound above is tight, we can take the graph shown in Figure 2. In Section 3 we will show other examples.
As shown in [18], the domination number of any graph of maximum degree is bounded below by . Therefore, the following result is deduced from Theorem 12.
Corollary 15**.**
For any graph of order and maximum degree ,
[TABLE]
The bound above is achieved, for instance, for any graph with , as in theses cases . An example of graph with and is the one shown in Figure 2.
By Theorems 5 and 12 we deduce the following result.
Theorem 16**.**
Let be a bipartite graph. If , then
[TABLE]
Theorem 33 will provide a family of Cartesian product graphs where
The line graph of a simple non-empty graph is obtained by associating a vertex with each edge of and connecting two vertices of with an edge if and only if the corresponding edges of have a vertex in common.
Theorem 17**.**
For any graph of order ,
[TABLE]
Proof.
Let be a -packing of such that . Let be two disjoint sets of cardinality such that and for every , i.e., every edge in has an endpoint in and the other one in . Since is a -packing of , for any we have . Thus, is a super dominating set of , as for every there exists such that . Hence,
[TABLE]
as required. ∎
To show that the bound above is tight we can take, for instance, both graphs shown in Figure 3. Notice that in these cases Theorem 17 gives a better result than Corollary 14.
3 Super domination in Corona product graphs
The corona product graph is defined as the graph obtained from and by taking one copy of and copies of and joining by an edge each vertex from the copy of with the vertex of [6]. It is readily seen that . Therefore, The bounds obtained in Theorem 12 and Corollary 14 are tight, as for the corona graph or , where is an arbitrary graph of order and .
Theorem 18**.**
For any graph of order and any nonempty graph
[TABLE]
Furthermore, for any integer ,
[TABLE]
Proof.
We first assume that is nonempty. Let be the order of , the vertex set of , and the vertex set of the copy of associated to . Let be a -set and the set associated to in the copy of . It is readily seen that is a super dominating set of . Thus,
[TABLE]
Now, let be a -set. If , for some , then , which implies that is a super dominating set of and , as . Thus, from now on we can assume that is taken in such a way that . Now, let . If there exists such that , then set of vertices of associated to is a super dominating set of and , which is a contradiction. Hence, for every , which implies that
[TABLE]
Therefore, the first equality holds.
On the other hand, Theorems 3 and 12 imply that . ∎
An alternative proof for the result above can be derived from a formula obtained in [13] for the super domination number of rooted product graphs. We leave the details to the reader.
4 Super domination in Cartesian product graphs
The Cartesian product of two graphs and is the graph whose vertex set is and two vertices are adjacent in if and only if either
- •
and , or
- •
and .
The Cartesian product is a straightforward and natural construction, and is in many respects the simplest graph product [7, 11]. Hamming graphs, which includes hypercubes, grid graphs and torus graphs are some particular cases of this product. The Hamming graph is the Cartesian product of copies of the complete graph . Hypercube is defined as . Moreover, the grid graph is the Cartesian product of the paths and , the cylinder graph is the Cartesian product of the cycle and the path , and the torus graph is the Cartesian product of the cycles and .
This operation is commutative in the sense that , and is also associative in the sense that . A Cartesian product graph is connected if and only if both of its factors are connected.
This product has been extensively investigated from various perspectives. For instance, the most popular open problem in the area of domination theory is known as Vizing’s conjecture. Vizing [17] suggested that the domination number of the Cartesian product of two graphs is at least as large as the product of domination numbers of its factors. Several researchers have worked on it, for instance, some partial results appears in [2, 7]. For more information on structure and properties of the Cartesian product of graphs we refer the reader to [7, 11].
Before obtaining our first result we need to introduce some additional notation. The set of all -sets will be denoted by . For any we define the set formed by subsets of cardinality such that for every there exists such that With this notation in mind we define the following parameter which will be useful to study the super domination number of Cartesian product graphs.
[TABLE]
With the aim of clarifying what this notation means, we consider the graphs shown in Figure 3. For the graph (on the left) we have , and , while for the graph (on the right) we have that , and . Notice that and .
If has vertices and , then is a universal vertex of . It is readily seen that the following remark holds.
Remark 19**.**
Let be a universal vertex of a graph of order and let . If for some , then
For instance, for the graph shown in Figure 3 we have , and .
Theorem 20**.**
For any graphs and of order and , respectively,
[TABLE]
Proof.
The lower bound is deduced from Theorem 1, so we proceed to deduce the upper bound. Let be a -set, and such that and . We claim that for any -set , the set
[TABLE]
is a super dominating set of . To see this we fix . Notice that or , so that we differentiate these two cases.
Case 1: . In this case, there exists such that . Since , and
[TABLE]
we can conclude that
Case 2: . In this case and . Since is a super dominating set of , there exists such that . Also, if there exists , then is a super dominating set of , which is a contradiction, so that is an independent set. Hence,
[TABLE]
Therefore, is a super dominating set of , which implies that
[TABLE]
as required. ∎
As a direct consequence of Theorem 20 we derive the following bound.
Corollary 21**.**
For any graphs and of order and , respectively,
[TABLE]
We will see in Theorem 30 that if , then , which implies that the bound given in Corollary 21 is tight, as for .
The following result is a direct consequence of Theorem 20 and Corollary 21.
Theorem 22**.**
Let and be two graphs of order and , respectively. If or , then
[TABLE]
From the result above we have that for any graph of order ,
[TABLE]
Since the hypercube graph is defined as , for , and , Corollary 22 leads to the following result.
Remark 23**.**
For any integer ,
[TABLE]
From Theorems 1 and 2, and Corollary 21 we deduce the following result.
Theorem 24**.**
Let be an integer. The following statements hold for any graph of order .
- •
If , then .
- •
If , then .
- •
If , then .
- •
If , then .
- •
If , then .
As usual in domination theory, when studying a domination parameter, we can ask if a Vizing-like conjecture can be proved or formulated.
Conjecture 25**.**
(Vizing-like conjecture)* For any graphs and ,*
[TABLE]
The above conjecture holds in the following case, which is a direct consequence of Theorem 22.
Remark 26**.**
Let and be two graphs of order and , respectively. If or , then
[TABLE]
In order to deduce another consequence of Theorem 20 we need to state the following lemma.
Lemma 27**.**
Let be the number of vertices of degree one of a graph , and let . If there exists a universal vertex of such that , for some , then .
Proof.
Let be a universal vertex of . If , then we are done, so that we assume that .
We first suppose that . If , then for any and the universal vertex of belongs to . So we assume that . In such a case, for any pair of adjacent vertices we have that and , which implies that for any vertex of degree one, . Hence, .
Now, suppose that . By Remark 19, for any and the universal vertex does not belong to , which implies that for any vertex of degree one, . Thus, . ∎
By Lemma 27 we can derive a consequence of Theorem 20 in which we replace the parameter by the number of vertices of degree one in .
Proposition 28**.**
Let be the number of vertices of degree one of a graph of order , and let . If there exists a universal vertex of such that , for some , then for any graph of order ,
[TABLE]
In order to see that the bound above is tight, we can observe that for and , , we have
Notice that, by Remark 19, a particular case of Proposition 28 can be stated as follows.
Corollary 29**.**
Let be the number of vertices of degree one of a graph of order and maximum degree . If , then for any graph of order ,
[TABLE]
We now provide a tight bound on in terms of the order of and .
Theorem 30**.**
For any nonempty graphs and of order and , respectively,
[TABLE]
Furthermore, for any integers and ,
[TABLE]
and for any integer ,
[TABLE]
Proof.
Let and such that and . Now, let such that
[TABLE]
To check that is a super dominating set of we only need to observe that for any there exists such that and for any there exists such that . Hence, .
To conclude the proof, it remains to consider the Cartesian product of complete graphs. Let be a -set. Notice that if , and , then leads to and leads to . Furthermore, if , then , as for any we have that and . Analogously, if , then . Hence,
[TABLE]
Thus, if and , then , which implies that .
Moreover, if , then Equation (5) leads to . To conclude that we only need to observe that for any the set is a super dominating set of . Therefore, the result follows. ∎
The independence number of any Cartesian product graph is bounded below as follows.
Theorem 31**.**
[16]* For any graphs and of order and , respectively,*
[TABLE]
From Theorems 6, 7 and 31 we deduce the following result.
Theorem 32**.**
For any pair of bipartite graphs and ,
[TABLE]
Theorem 33**.**
For any integers ,
[TABLE]
Proof.
By Theorem 32, . Next, we proceed to show that . Let , , and , , be the center and the set of leaves of and , respectively. Let and . Note that, for each vertex , there exists exactly one vertex such that . So, is a super dominating set of with cardinality , and thus . Therefore, . ∎
Acknowledgements
This research was supported in part by the Spanish government under the grants MTM2016-78227-C2-1-P and PRX17/00102.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] B. Bollobás, E. J. Cockayne, Graph-theoretic parameters concerning domination, independence, and irredundance, J. Graph Theory 3 (3) (1979) 241–249.
- 2[2] B. Brešar, P. Dorbec, W. Goddard, B. L. Hartnell, M. A. Henning, S. Klavžar, D. F. Rall, Vizing’s conjecture: a survey and recent results, Journal of Graph Theory 69 (1) (2012) 46–76.
- 3[3] E. J. Cockayne, P. J. P. Grobler, W. R. Gründlingh, J. Munganga, J. H. van Vuuren, Protection of a graph, Utilitas Mathematica 67 (2005) 19–32.
- 4[4] M. Dettlaff, M. Lemańska, J. A. Rodríguez-Velázquez, R. Zuazua, On the super domination number of lexicographic product graphs, Discrete Applied Mathematics. To appear. https://doi.org/10.1016/j.dam.2018.03.082
- 5[5] J. Egerváry, Matrixok kombinatorikus tulajdonságairól, Matematikai és Fizikai Lapok 38 (1931) 16–28.
- 6[6] R. Frucht, F. Harary, On the corona of two graphs, Aequationes Mathematicae 4 (3) (1970) 322–325.
- 7[7] R. Hammack, W. Imrich, S. Klavžar, Handbook of product graphs, Discrete Mathematics and its Applications, 2nd ed., CRC Press, 2011.
- 8[8] F. Harary, J. P. Hayes, H.-J. Wu, A survey of the theory of hypercube graphs, Comput. Math. Appl. 15 (4) (1988) 277–289.
