Available online at www.sciencedirect.

com

Electronic Notes in Discrete Mathematics 63 (2017) 77–84 www.elsevier.com/locate/endm

Complexity Issues of Variants of Secure

Domination in Graphs

Devendra Lad 1,2 P. Venkata Subba Reddy 3 J. Pavan Kumar 4

Computer Science and Engineering
NIT Warangal
Warangal, India

Abstract
Let G = (V, E) be a simple, undirected and connected graph. A connected dominat-
ing set S ⊆ V (G) is a secure connected dominating set of G, if for each u ∈ V (G)\S,
there exists v ∈ S such that (u, v) ∈ E(G) and the set (S \ {v}) ∪ {u} is a connected
dominating set of G. The minimum cardinality of a secure connected dominating
set of G denoted γsc (G), is called the secure connected domination number of G. In
this paper, we show that the decision problems of finding a minimum secure con-
nected dominating set and a minimum secure total dominating set are NP-complete
for split graphs. We initiate the study of 2-secure domination and show that the
decision problem corresponding to the computation of 2-secure domination number
of a graph is NP-complete, even when restricted to split graphs or bipartite graphs.
Finally we show that given two positive integers k (≥ 2) and n (≥ max{4, k}) there
exists a graph G with |V (G)| = n and γsc (G) = k.
Keywords: Secure dominating set, split graph, NP-complete, total domination.

1
Thanks to everyone who should be thanked
2
Email: deepak.devendra@gmail.com
3
Email: pvsr@nitw.ac.in
4
Email: jp@nitw.ac.in

https://doi.org/10.1016/j.endm.2017.11.001
1571-0653/© 2017 Elsevier B.V. All rights reserved.
78 D. Lad et al. / Electronic Notes in Discrete Mathematics 63 (2017) 77–84

1 Introduction
Let G(V, E) be a simple, undirected and connected graph. We use commonly
used terminology and notations as found in popularly available texts such as
[2,7]. For a vertex v ∈ V (G), the (open) neighborhood of v in G is N (v)=
{u ∈ V (G) : (u, v) ∈ E(G)}, the closed neighborhood of v is defined as N [v] =
N (v) ∪ {v}. If S ⊆ V (G), then the (open) neighborhood of S is the set
N (S) = ∪v∈S N (v). The closed neighborhood of S is N [S] = S ∪ N (S). Let
S ⊆ V (G) and v be a vertex in S. The S-external private neighborhood of v
denoted epn[v, S] is defined as {w ∈ V (G)|N (w) ∩ S = {v}}.
A subset S of V (G) is a dominating set (DS) in G if for every u ∈ V (G) \ S,
there exists v ∈ S such that (u, v) ∈ E(G), i.e., N [S] = V (G). The domination
number of G is the minimum cardinality of a DS in G and is denoted by γ(G).
A set S ⊆ V (G) is said to be a secure dominating set (SDS) in G if for every
u ∈ V (G) \ S there exists v ∈ S such that (u, v) ∈ E(G) and (S \ {v}) ∪ {u}
is a dominating set. The minimum cardinality of a SDS in G is called the
secure domination number of G and is denoted by γs (G). A dominating set S
is said to be a connected dominating set (CDS), if the induced subgraph G[S]
is connected. Let S be a CDS in G. A CDS S is called a secure connected
dominating set (SCDS) in G, if for each u ∈ V (G) \ S, there exists v ∈ S such
that (u, v) ∈ E(G) and (S \ {v}) ∪ {u} is a CDS in G. The secure connected
domination number γsc (G) of G is the smallest cardinality of a SCDS in G.
A dominating set S is said to be a total dominating set (TDS), if the induced
subgraph G[S] has no isolated vertices. A TDS S is called a secure total
dominating set (STDS) of G, if for each u ∈ V (G) \ S, there exists v ∈ S
such that (u, v) ∈ E(G) and (S \ {v}) ∪ {u} is a TDS in G. The secure total
domination number γsc (G) of G is the smallest cardinality of a STDS in G.

2 Complexity of Secure Domination Parameters

In this section we determine the algorithmic complexity of different secure
domination parameters when restricted to split or bipartite graphs.

2.1 Algorithmic complexity of secure connected (total) domination in split

graphs
First we recall the following definition of split graph.
Definition 2.1 A graph is a split if its vertex set is the disjoint union of a
clique and independent set.
 D. Lad et al. / Electronic Notes in Discrete Mathematics 63 (2017) 77–84 79

Now, we state the theorem which has been proved in [1] with the following
definitions of leaf and support vertices. A leaf u of a graph G is a vertex of
degree one and the support vertex of a leaf u is the unique vertex v such that
(u, v) ∈ E(G). We denote by L(G) and S(G) be the set of leaves and support
vertices of G, respectively.
Theorem 2.2 Let G be a connected graph of order n ≥ 3 and let X be a
secure connected dominating set of G. Then
(i) L(G) ⊆ X and S(G) ⊆ X
(ii) No vertex in L(G) ∪ S(G) is an X-def ender .
The following decision problem for the connected domination number of a
graph is known to be NP-complete, even when restricted to split graphs [4].
Connected Domination Problem (CDOM)
Instance: An undirected and connected graph G and a positive integer k.
Question: Is γc (G) ≤ k ?
Formulated as a decision problem, the secure connected domination problem,
denoted by SCDOM, is defined as follows:
Secure Connected Domination Problem (SCDOM)
Instance: An undirected and connected graph G and a positive integer k.
Question: Is γsc (G) ≤ k ?
Theorem 2.3 SCDOM is NP-complete for split graphs
Proof. Clearly, the SCDOM is a member of NP, since we can check in poly-
nomial time whether or not a set of vertices is SCDS of G. Now let us show
how a polynomial time algorithm for SCDOM could be used to solve CDOM
in polynomial time. Given a positive integer k and split graph G whose vertex
set is partitioned into a clique Q and an independent set I, we construct a
graph G∗ with clique Q∗ and independent set I ∗ as follows.
V (G∗ ) = V (G) ∪ {x, y}, where x and y are newly introduced vertices,
E(G∗ ) = E(G) ∪ {(x, u) : ∀u ∈ V (G)} ∪ (x, y).
It is worth pointing out that G∗ is a split graph, where V (Q∗ ) = V (Q) ∪ {x}
and V (I ∗ ) = V (I) ∪ {y}. We also note that |V (G∗ )| = |V (G)| + 2 and
|E(G∗ )| = |E(G)| + |V (G)| + 1, and so G∗ can be constructed from G in poly-
nomial time. Next, we shall show that G has a CDS of size at most k if and
only if G∗ has a SCDS of size at most k + 2.
Let S be a CDS of size at most k in G. Then it is evident from Theorem 2.2
that S ∪ {x, y} is a SCDS of G∗ of cardinality k + 2 .
Conversely, suppose that G∗ has a SCDS S ∗ of size at most k + 2. Let
 
S = S ∗ ∩ V (G). It is clear that |S ∗ ∩ {x, y}| = 2 and hence |S | ≤ k. Let
80 D. Lad et al. / Electronic Notes in Discrete Mathematics 63 (2017) 77–84


w ∈ V (G)\S . Now if S ∗ becomes (S ∗ \{x})∪{w} then y is an isolated vertex

and it is no longer a CDS. Thus there exists v ∈ S such that (S ∗ \ {v}) ∪ {w}

is a CDS. Hence S is a SDS in G. Further since G is a split graph each vertex
 
u ∈ {S ∩ V (I)} can be replaced by a vertex v ∈ {N [u] ∩ V (Q)}. Thus S is
 
a SCDS in G and |S | ≤ k. Every SCDS is a CDS and hence S is a CDS of
size at most k in G. 2
The following decision problem for the total domination number of a graph is
known to be NP-complete, even when restricted to split graphs [5].
Total Domination Problem (TDOM).
Instance: An undirected and connected graph G and a positive integer k.
Question: Is γt (G) ≤ k ?
Formulated as a decision problem, the secure total domination problem, de-
noted by STDOM is defined as follows:
Secure Total Domination Problem (STDOM).
Instance: An undirected and connected graph G and a positive integer k.
Question: Is γst (G) ≤ k ?
Theorem 2.4 STDOM is NP-complete for split graphs.
Proof. A SDS is a STDS if it contains no isolated vertices. The proof used
to show that the SCDOM is NP-complete can be used to show that STDOM
is NP-complete. 2

2.2 Algorithmic complexity of 2-secure domination in split and bipartite graphs

First we recall the definition of secure domination in graph.
Definition 2.5 If S is a dominating set in G then vertex v ∈ S defends
u ∈ V \ S and u ∈ N (v) if (S \ {v}) ∪ {u} is a dominating set in G. It is
denoted as v S-defends u [3].
Definition 2.6 If S is a secure dominating set in G then for each u ∈ V \ S
there exists a vertex v ∈ S such that v S-defends u [3].
We initiate the study of 2-secure domination in graphs with the following
definition.
Definition 2.7 A set S ⊆ V (G) is called a 2-secure dominating set (2-SDS)
in G if for every pair of vertices u1 , u2 ∈ V (G) there exists a pair of vertices
v1 , v2 ∈ S, v1 ∈ N [u1 ] and v2 ∈ N [u2 ] such that (S \ {v1 , v2 }) ∪ {u1 , u2 } is a
dominating set in G. The 2-secure domination number denoted γ2s (G) equals
the minimum cardinality of a 2-SDS in G.
 D. Lad et al. / Electronic Notes in Discrete Mathematics 63 (2017) 77–84 81

Lemma 2.8 If S is a 2-SDS of a graph G then for each u ∈ V \ S there exists

at least two vertices {w, v} ∈ S such that v S-defends u and w S-defends u.

Proof. Proof follows from the definition of 2-SDS. 2

The following decision problem for the secure domination number of a graph is
known to be NP-complete, even when restricted to split graphs and bipartite
graphs citeosd.
Secure Domination Problem (SDOM)
Instance: An undirected and connected graph G and a positive integer k.
Question: Is γs (G) ≤ k ?
Formulated as a decision problem, the 2-secure domination problem, denoted
by 2-SDS is defined as follows:
2-Secure Domination Problem (2-SDOM)
Instance: An undirected and connected graph G and a positive integer k.
Question: Is γ2s (G) ≤ k ?
Theorem 2.9 2-SDOM is NP-complete for split graphs.

Proof. Clearly, 2-SDOM is a member of NP, since we can check in polynomial

time whether or not a given set of vertices is a 2-SDS of G. Now let us show
how a polynomial time algorithm for 2-SDOM could be used to solve SDOM in
polynomial time. Given a positive integer k and a split graph G whose vertex
set is partitioned into a clique Q and an independent set I, we construct a
graph G∗ with clique Q∗ and independent set I ∗ as follows.
V (G∗ ) = V (G) ∪ {x, y}, where x and y are newly introduced vertices,
E(G∗ ) = E(G) ∪ {(x, u) : ∀u ∈ V (G)} ∪ (x, y).
It is worth pointing out that G∗ is a split graph, where V (Q∗ ) = V (Q) ∪ {x}
and V (I ∗ ) = V (I) ∪ {y}. We also note that |V (G∗ )| = |V (G)| + 2 and
|E(G∗ )| = |E(G)| + |V (G)| + 1, and so G∗ can be constructed from G in
polynomial time.Next, we shall show that G has a SDS of size at most k if
and only if G∗ has a 2-SDS of size at most k + 2. Let S be a SDS of size at
most k. Then it is evident that S ∗ = S ∪ {x, y} is a 2-SDS of G∗ of cardinality
at most k + 2. Conversely, suppose that G∗ has a 2-SDS S ∗ with |S ∗ | ≤ k + 2.
By lemma 2.8, it is evident that y is part of every 2-SDS. Let x is not part
of some 2-SDS of G∗ , then we can give a 2-SDS of same cardinality which
includes x, since x is adjacent to all vertices of G∗ . Hence {x, y} is part of

every 2-SDS of size at most k + 2. Let S = S ∗ ∩ V (G). It is clear that
 
|S ∗ ∩ {x, y}| = 2 and hence |S | ≤ k. If S securely dominates G then we are

done. Suppose now that S is not a SDS of G. Let A be the set of vertices of
82 D. Lad et al. / Electronic Notes in Discrete Mathematics 63 (2017) 77–84

 
G having no vertex S -defending it in S and let w ∈ A. Since S ∗ is a 2-SDS
there must exist at least two vertices which S ∗ -defends w in G∗ by lemma 2.8.

But NG∗ [w] ∩ (S ∗ \ S ) = {x} i.e., only single vertex x S ∗ -defends w in G∗ ,
hence the contradiction. Therefore A is an empty set. Hence G has a SDS of
size at most k. 2

Theorem 2.10 2-SDOM is NP-complete for bipartite graphs.

Proof. Clearly, 2-SDOM is a member of NP. We next show how a polynomial

time algorithm for 2-SDOM could be used to solve SDOM in polynomial time.
Given a bipartite graph G with partite sets X and Y , and a positive integer
k, we construct a bipartite graph G∗ with partite sets X ∗ and Y ∗ as follows.
V (G∗ ) = V (G) ∪ {x1 , x2 , y1 , y2 }, where x1 , x2 , y1 , y2 are newly introduced
vertices,
E(G∗ ) = E(G)∪{(x1 , x2 ), (y1 , y2 )}∪{(x1 , u) : ∀u ∈ X}∪{(y1 , u) : ∀u ∈ Y }.
It is worth pointing out that G∗ is a bipartite graph. |V (G∗ )| = |V (G)| + 4
and |E(G∗ )| = |E(G)| + |V (G)| + 2, and so G∗ can be constructed from G in
polynomial time.
We will show that G has a SDS of size at most k if and only if G∗ has a 2-SDS
of size at most k + 4. Let S be a SDS of G of size at most k. Then clearly
S ∗ = S ∪ {x1 , y1 , x2 , y2 } is a 2-SDS of G∗ of size at most k + 4. Conversely,
suppose that G∗ has a 2-SDS S ∗ of size at most k + 4. By lemma 2.8, it is
clear that {x2 , y2 } is part of every 2-SDS. Let {x1 , y1 } is not part of some
2-SDS of G∗ , then we can give a 2-SDS of same cardinality which includes
{x1 , y1 }, since x1 is adjacent to all vertices of X ∗ and y1 is adjacent to all
vertices of Y ∗ . Hence {x1 , x2 , y1 , y2 } is part of every 2-SDS of size at most

k + 4. Let S = S ∗ ∩ V (G). It is clear that |S ∗ ∩ {x1 , x2 }| = 2 and likewise
 
|S ∗ ∩ {y1 , y2 }| = 2. Hence S ≤ k. Now if S securely dominates G, then we
are done.

Let us assume that S is not a SDS in G. Let B ⊂ V (G) be the set of vertices

not securely dominated by S , and let Bx = B ∩ X and By = B ∩ Y . We
assume that Bx
= ∅. Since Bx is 2-securely dominated by S ∗ in G∗ , therefore
there exist at least two vertices which S ∗ -defends u ∈ Bx in G∗ by lemma 2.8.
But NG∗ [u] ∩ (X ∗ \ X) = {x1 } i.e., only single vertex x S ∗ -defends w in G∗ ,
hence the contradiction. Hence Bx is an empty set. Similarly we can prove

By is an empty set. Hence S is a SDS of size at most k. 2
 D. Lad et al. / Electronic Notes in Discrete Mathematics 63 (2017) 77–84 83

3 Constructing Graphs with given Order and γsc (G)

In this section we show that given two positive integers k (≥ 2) and n (≥
max{4, k}) there exists a graph G with |V (G)| = n and γsc (G) = k.

Theorem 3.1 Let n (≥ 4) and k (≥ 2) be two positive integers. For any

n ≥ k, there exists a simple, undirected and connected graph G with n vertices
such that γsc (G) = k.

Proof. We form the graph G such that V (G) = {v1 , v2 , v3 , ..., vn } and E(G) =
{(vi , vi+1 ) | 1 ≤ i ≤ k − 1} ∪ {(v1 , vj ) | k + 1 ≤ j ≤ n} ∪ {(vk , vj ) | k + 1 ≤
j ≤ n}. Let Vp = {v1 , v2 , v3 , ..., vk }, Vq = {vk+1 , vk+2 , ..., vn }, Ve = {v1 , vk },
V2e = {v2 , vk−1 } , Vwe = {v2 , v3 , v4 , ..., vk−1 } and Vw2e = Vwe \ V2e .
Let S be any SCDS of graph G such that |S| = k. From [1] we know that
γsc (Pn ) = n for all n ≥ 3.
Case 1 n = k.
Clearly G ∼
= Pk hence the result.
Case 2 n = k + 1.
As G ∼
= Ck+1 the result follows from [1].
Case 3 n > k + 1.
In this case we also show that γsc (G) = k. First it can be easily seen that Vp
is a SCDS of G. Hence γsc (G) ≤ k. Now we show that γsc (G) ≥ k.
Claim 1: Every SCDS of G must contain the two vertices in the set Ve .
Proof : Let R be any SCDS If we remove v1 (or vk ) from the set R then vk (or
v1 ) will have n − k − 1 external private neighbours and it is no longer SCDS
[1]. Hence the claim.
Claim 2: Vw2e is part of every SCDS.
Proof : Let S ⊆ V (G) of cardinality at least k such that it does not contain
at least one vertex v ∈ Vw2e . Now, there exists a vertex w ∈ Vq \ S which will
not be securely defended and set S is not a SCDS. Hence, Vw2e is part of every
SCDS.
Claim 3: At least one vertex from V2e should be a part of every SCDS.
Proof : Let us prove it by contradiction, by assuming there exists some SCDS
S with no vertex from V2e . However, by our assumption induced subgraph
G[S] is no longer connected to any vertex of Vw2e , which is a contradiction.
If only one vertex from V2e is included in any SCDS S, then S should contain
at least one vertex from Vq , otherwise G[S] is not even connected. From claim
84 D. Lad et al. / Electronic Notes in Discrete Mathematics 63 (2017) 77–84

1,2 and 3 we say that γsc (G) ≥ k. Hence γsc (G) = k. Therefore there exists a
graph G = (V, E) such that γsc (G) = k and |V | = n. 2

4 Conclusion
In this paper, we have shown that secure connected (total) dominating set and
2-secure dominating set problems are NP-complete for split graphs. We have
also shown that 2-secure domination problem is NP-complete for bipartite
graphs. Determining computational complexity of secure connected (total)
domination for bipartite graphs is still open. It will be interesting to see
if same results can be obtained for secure connected (total) domination for
chordal graphs and circle graphs.
Acknowledgements
We are grateful to Prof. S. Arumugam, Director, n-CARDMATH and
Senior Professor (Research), Kalasalingam University, for introducing us to
the concept of secure domination in graphs and suggesting its generalizations.

5 Bibliographical references

References

[1] Cababro, Amerkhan G., Sergio S. Canoy, and Imelda S. Aniversario, Secure
connected domination in a graph, International Journal of Mathematical
Analysis 8 (2014), 2065-2074.
[2] Chartrand, G., and P. Zhang, “Introduction to Graph Theory,” Tata McGraw-
Hill, New Delhi, (2005).
[3] Cockayne, E.J, P.J.P. Grobler, W.R. Grundlingh, J. Munganga, and J.H. van
Vuuren, Protection of a graph, Util. Math. 67 (2005), 1932.
[4] Guha, S., and S. Khuller, Approximation algorithms for connected dominating
sets, Algorithmica 20 (1988), 374-387.
[5] Laskar, R.C., and J. Pfaff , ”Domination and irredundance in split graphs,”
Technical Report 430, Clemson University, Dept. Math. Sciences, (1983).
[6] Merouane, H.B.,and Mustapha Chellali, On secure domination in graphs,
Information Processing Letters 1150 (2015), 786-790.
[7] West, D.B., “An Introduction to Graph Theory,” 2nd Ed., Prentice-Hall, N.J.,
(1946).