International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169

Volume: 5 Issue: 8 164 – 174

_______________________________________________________________________________________________

Some Properties on Strong Roman Domination in Graphs

G. Suresh Singh#1, Narges Mohsenitonekboni#2

#
Department of Mathematics
University of Kerala
Kariyavattom-695581
Thiruvananthapuram-Kerala, India
1
sureshsinghg@yahoo.co.in
2
narges.mtonekabni@gmail.com

Abstract—A Strong Roman dominating function (SRDF) is a function f : V   0 ,1, 2 , 3 satisfying the condition that every
vertex u for which f u   0 is adjacent to at least one vertex v for which f  v   3 and every vertex u for which f u   1 is
adjacent to at least one vertex v for which f  v   2 . The weight of an SRDF is the value f  V    uV f u  . The minimum weight
of an SRDF on a graph G G . In this paper, we attempt to verify some properties on SRDF
is called the Strong Roman domination numberof
and moreover we present Strong Roman domination number for some special classes of graphs. Also we show that for a tree T with n  3
6n  l  s
vertices, l leaves and s support vertices, we have  SR  T   and we characterize all trees achieving this bound.
4

Keywords-Roman domination number, Strong Roman domination number, Graph, Tree, Star, Double star, Connected graph.

__________________________________________________*****_________________________________________________

I. INTRODUCTION literature) is a vertex of degree at least 2 that is adjacent to at

Mathematical study of domination in graphs began
around 1960, there are some references to domination related
problems about 100 years prior. In 1862, De Jaenisch [2]
attempted to determine the minimum number of queens
required to cover a n  n chess board. Except as indicated
otherwise, all terminology and notation follows [5, 4, 9]. Let
G  V , E  be a graph of order V  n . For any vertex
v  V the open neighborhood of v is the set
least one leaf. A support vertex adjacent to two or more leaves
is a Strong support vertex. A Weak support vertex is a support
vertex that is adjacent to exactly one leaf. Also we denote the
 
set of leaves in G by L G and the setof support vertices
by S(G). A Star is the graph K1 , k where k  1 . If k  1 , the
vertex of degree k is called the Center vertex of the star. A
Double star is formed from two disjoint stars by joining the
center vertices of each by an edge. Thus a Double star is a tree

N  v   u  V uv  E and the closed neighborhood is

the set N v   N  v   v  . For a set S  V the open
neighborhood of S is N  S    N  v  and the closed
vS
neighborhood is N  S   N  S   S .A set S of vertices is
called a vertex cover if for every edge u v  E either u  S
or v  S . A graph G is said to be connected if there is at
least one path between every pair of vertices in G . Otherwise,
G is disconnected. A graph with no cycle is acyclic. A forest
is an acyclic graph. A tree is a connected acyclic graph. A
rooted tree T distinguishes a vertex r called the root. A
vertex of degree 1 is called a leaf which denoted by l . A
adjacent leafof vertex u in a tree T is a neighborhood of u
that is a leaf in T . A support vertex (also called a stem in the
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
with exactly two vertices that are not leaves.
We now introduce the concept of dominating sets in
graphs. A set S  V is a dominating set if N S  V or  
equivalently, every vertex in V  S is adjacent to at least one
vertex in S . The domination number  G  is the minimum
cardinality of a dominating set in G and a dominating set S
of minimum cardinality is called a   G   set of G , see
[10]. Let f : V   0 ,1, 2  be a function having the
property that for every vertex v  V with f  v   0 , there
exists a neighborhood u  N  v  with f u   2 . Such a
function is called a Roman dominating function or just an
RDF. The weight of an RDF is the value
f  V    uV f u  . The minimum weight of an RDF on
164
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
G is called the Roman domination number of G and is It is implied that V2  V3 is a dominating set of G . So
denoted by  R  G  , see [1, 10].   G   V2  V3 , thus
K. Selvakumar et al. [8] introduced Strong Roman
domination in 2016. A Strong Roman dominating function 2   G   2 V2  2 V3
 
(SRDF) is a function f : V  0 , 1 , 2 , 3 satisfying the  V1  2 V2  3 V3
u for which f  u   0 is
  SR  G .
condition that every vertex
adjacent to at least one vertex v for which f  v   3 and Hence
every vertex u for which f  u   1 is adjacent to at least 2   G    SR  G . (1)
one vertex v for which f  v   2 . The weight of an SRDF Now, let S be a   set of G . Then   G   S . We can
is the value f  V    f u  . The minimum weight of define an SRDFon G , for all v  S we have f  v   3 and
uV

an SRDF on a graph G is called the Strong Roman also for all u  S we have f  u   0 . Therefore
domination numberof G .
In 2004, Cockayne et al. [1] studied the graph
 V0 , V1 , V2 , V3     ,  ,  , S  is an SRDF. It is
theoretic properties of Roman dominating sets. In recent years impliedthat V0  0 , V1  0 , V2  0 and V3  S
several authors studied the concept of Roman dominating . Therefore
 SR  G   3 V3
functions and Roman domination numbers [12, 6, 7, 11]. In
this paper, we present some results on SRDF and Strong
Roman domination number for some special classes of graphs.
3 S
Also we show that for a tree T with n  3 vertices, l leaves
6n  l  s  3   G .
and s support vertices,  SR  T   and we Hence
 SR  G   3  G  .
4
characterize all trees achieving this bound. (2)
From (1) and (2) we get 2   G    SR  G   3   G  . □
Proposition: For any graph G , there exists an SRDF,
f  V0 ,V1 ,V2 ,V3  of G , such that V1   . Theorem 2: For any graph G of order n ,
 SR G   2  G 
Proof:
if and only if G  K n .
Let V1   and u  V . By the definition of SRDF,
there exists a vertex v  V2 such that v  N  u  . Hence a Proof:
function g  V0  u  , V1  u  , V2  v  , V3  v  is Suppose that f  V0 ,V1 ,V2 ,V3  isa
an SRDF. Continuing with the same argument we find an  SR  G   function . Thus V2  V3 is a dominating set of
SRDF with V1   . Therefore, the proposition follows. □
the graph G . Therefore  G   V2  V3 . The equality
Theorem 1: For any graph G,  SR G   2  G  implies that we have equality in
2   G    SR  G   3  G . 2   G   2 V2  2 V3

Proof:  2 V2  3 V3
Suppose that f   V0 , V1 , V2 , V3  is a   SR  G .
 SR  G   function and V0  n0 , V1  n1 , So V3  0 , which implies that V0   . Hence all vertices
V2  n2 and V3  n3 . are assigned with 2 and therefore

 SR  G   f  V
  SR  G   2 V2
  f u   2 n.
u V This implies that  G   n which shows that G  K n .
 3 n3  2 n2  n1 . Conversely, It is obvious that if G  K n , then
V3  V0  The set V3 dominates the set V0 .  SR G   2  G  . □
V2  V1  The set V2 dominates the set V1 .
165
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
Theorem 3: For any graph G,  R  G   1  V2  1
 R  G    SR  G  .
 2 V2  1
Proof:   SR  G   1 .
Suppose that f  V0 ,V1 ,V2 ,V3  is a Thus  R  G   1   SR  G  which is a contradiction.
 SR  G   function . Let g   Y0 , Y1 , Y2  be an RDF Therefore in this case, V2  1 and thus G  K1 .
on G where Y0  V0 , Y1  V2 and Y2  V3 . Therefore
Now, assume that V2  0 and if V3  1 , then
 R  G   Y1  2 Y2
 R  G   2 V3
 V2  2 V3
 3 V3  2 .
 2 V2  3 V3 Similarly, we get  R  G   1   SR  G  , which is a
  SR  G  . contradiction. Therefore in this case, V3  1 and if
 R  G    SR  G  . □
V3   v , then deg  v   n  1 . Thus G has a vertex of
Thus

Based on above theorem, we know that degree n  1 .

 SR  G    R  G   1 . In the next theorem, we will Therefore in each case   G   n  1 .
discuss the equation of this inequality. Conversely, If   G   n  1 , then  SR  G   3

Theorem  SR  G    R  G   1
4: if and only if
and  R  G   2 . So  SR  G    R  G   1 . □
  G   n  1. Theorem 5: For any path Pn ,
n , n  0  mod 3 
Proof:
 SR  Pn   
 SR  G    R  G   1 and  n  1 , n  0  mod 3 
.
Suppose that
f   V0 , V1 , V2 , V3  isa  SR  G   function . Define
g   Y0 , Y1 , Y2  is an RDF on G where Y0  V0 ,
Proof:
Suppose that a , b and c are consecutive vertices
Y1  V2 and Y2  V3 . Therefore and f   V0 , V1 , V2 , V3  is a  SR  G   function of
 R  G   Y1  2 Y2 Pn , respectively. If two vertices of  a , b , c belongingto
 V2  2 V3 . V0 , then either one of those vertices belongs to V3 , which in
this case we have
V2  0 and also V3  0 , then
f a   f b   f c  3,
On the other hand, if
V2  2 V3  V2  2 V3   V2  1    V3  1 
or a , c  V0 and b  V2 . In this case, all vertices which are
 2 V2  3 V3  2 . adjacent to a and c are named x and y should belongto V3
. Therefore
f  x   f a   f b   f c   f  y  8.
Hence
 R  G   1  V2  2 V3  1 So, always
 2 V2  3 V3  1  SR  Pn   f  V 
  SR  G   1 .  n.
Thus  R  G   1   SR  G  , which is a contradiction. Now, we use an induction on the order n . Assume the result
is true for n  6 . Suppose that n  7 and it is true for
V3  0 or V2  0 .
m  n . If n  0  mod 3 and Pn  v1 v2  vn , we put
Therefore

Let V3  0 and if V2  1 , then , i  2  mod 3 

3
 R  G   V2 f  vi    .
0 , i  2  mod 3 
 2 V2  2 . Hence, f is an SRDF and also
Hence

166
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
n
g1  V    SR  Pi  1 
f  V    f  vi 
i 1  i 1,
n. g 2  V    SR  Pn  i 
Therefore  SR  Pn   n . ni.
On the other hand, we have shown for any n ,  SR  Pn   n . Therefore
Hence  SR  Pn   n .  SR  Pn   f  V 
Now, let n   0  mod 3  and f   V0 , V1 , V2 , V3  be a  g1  V   g 2  V   2
 SR  G   function .  i 1  n  i   2
If V2   , then it is easy to show that  n  1.
 SR  Pn   f  V   n  1. Hence  SR  Pn   n  1 .

If V2   , then assume that for 1  i  n , f  vi   2 . We Therefore in both cases  SR  Pn   n  1 .

consider the following cases: Now we define the function f on path Pn if
n  0  mod 3  as follows:
Case 1: i  1 or i  n.

Without loss of generality, suppose that i  1 ,

1) If n  1  mod 3  , then
f  v1   2 . Hence 3 , i  2  mod 3 

Pn  v1  Pn  1 . f  vi   0 , i  2  mod 3  , i  n .
2 , in
By the induction, we know that 
 SR  Pn 1   n  1 . On the other hand, it is
clearly, g f be an SRDFon Pn  1 . Thus
2) If n  2  mod 3  , then
, i  2  mod 3 
Pn  1
3
g  V    SR  Pn  1  f  vi    .
0 , i  2  mod 3 
 n  1.
Therefore In both cases f is an SRDF on path Pn of weight n  1 .
 SR  Pn   f  V  Hence  SR  Pn   n  1 .
 g V   2
Since we have shown that before  SR  Pn   n  1 ,
 n 1  2
therefore n  0  mod 3 
if , then we have
 n  1.  SR  Pn   n  1 . □
Hence  SR  Pn   n  1 .
Similarly, we can prove that the result is true for II. A NEW UPPER BOUND IN TREES
i  n.
It has been shown that the domination number of a
Case 2: i  1 , n . n
connected graph G of order n is at most [4]. Regarding
2
Hence we put Pi  1  v1 v2  vi  1 and the fact that  SR  G   3  G  , we get
Pn  i  vi  1 vi  2  vn . We know that
 SR  G   3   G  
3n
 SR  Pi  1   i  1 ,
.
2
 SR  Pn  i   n  i .
Our aim in this section is improve this bound on trees. We
show that for any tree T of order n with l leaves and s

  6 n  l  s . Moreover,
On the other hand, it is clearly, the functions
support vertices, we have  SR  T
g1  f Pi  1
and g2  f Pn  i
are two 4
we characterize all trees achieving this bound.
SRDFs on Pi  1 and Pn  i , respectively. Thus

167
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
Theorem 6: If T is a tree of order n  3 with l leaves and  
So, we can assume that diam T  4 . If T has a Strong
s support vertices, then support vertex u and also v and w are adjacent leaves to u ,
6n  l  s then we consider T   T  w . Let n  , l  and s  be order ,
 SR  T   .
4 number of leaves and number of vertices of tree T  ,
 
respectively. Since diam T  4 , we get n  3 .
Proof:
6 n  l   s 
We prove this by induction on order n of tree T . Therefore by induction  SR  T    . Suppose
If diam  T   2 , then T is a Star. Therefore l  n  1 , 4
that g   V0 ,  , V2 , V3  is a  SR  T    function .
s  1 and  SR  T   3 . Hence
If g  u   3 , then extension of g by assigning the weight
 SR  T   3
0 to w is an SRDF on tree T . Thus, since l   l  1 and
6n  l  s
 s   s , we have
 SR  T   g  V 
4
6 n   n  1  1
   SR  T  
4
6 n  l   s 

5n
. 
4 4
Now, assume that diam  T   3 . In this case, T is a 6  n  1    l  1  s

Double Star S a , b with central vertices u and v with degrees 4
6n  l  s
of a and b , respectively. Without loss of generality, suppose  .
that a  b . 4
Now, let g  u    3 . Then g  v   2 . Therefore the
If a  2 , then b  2 and thus T  P4 . Therefore
function f with f  w   f  v   0 , f  u   3 and for
 SR  T   5
any other vertex x , we have f  x   g  x  is an SRDF on
6n  l  s
 . tree T . Hence
4  SR  T   f  V 
Now, let a  3 . If b  2 , then the function
 g V   1
f   N  u  ,  , N  v  , u   is a
 SR  T   function . Since n  5 , l  n  2 and s  2 ,   SR  T    1
6 n  l   s 
we have
 1
 SR  T   f  V  4
 2 V2  3 V3 6  n  1    l  1  s
 1
 2 N v   3
4
6n  l  s
5  .
4
6n  l  s Therefore, we can consider the following Fact.
 .
4
Now, let b  3 . In this case, n  6 , l  n  2 , s  2 and Fact: T has no Strong support vertex.
the function f   V  T   u , v  ,  ,  , u , v   is an
SRDF on tree T . Hence We root the tree T at vertex x 0 . Support that
 SR  T   f  V  P  x0 x1  xd is a diagonal path. Based on Fact,
 2 V2  3 V3 deg  xd  1   2 . We consider the following cases:
6
6n  l  s
 
Case 1: deg xd  2  3 .
 .
4

168
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
In this case, every child of x d  2 is either a leaf or a child v other than xd  1 , then since
support vertex of degree 2 . Since based on Fact, T f   xd  2   0 , we can assume f   v   3 ,
does not has a Strong support vertex, we consider

T   T  xd , xd  1 
. So, n  n  2 , f   u   2 and the child weight of v is equals to
0 . By changing the weight of the vertices u and v
l   l  1 and s   s  1 . Suppose that
to 0 , x d  2 to 3 and child of v to 2 a new
f    V0 , V1 , V2 , V3 isa
 SR  T    function
 SR  T    function .
is obtained where weight of

f   xd  2   2
x d  2 is not equals to 0 , which is a contradiction.
If , then the function
Since we assume that for each
f   V0 , V1 , V2 , V3  where  SR  T    function , we have the weight of
V0  V0   xd , xd  2  , V1  V1   , x d  2 is 0 . So, we can assume that the only support
V2  V2   xd  2  and V3  V3   xd  is an child x d  2 is the vertex x d  1 . We put
SRDF on tree T . Thus
T   T  Txd  2 . In this case, n  n  4 . Since
 SR  T   f  V 
diam  T   4 , we get n  2 .
 f V   1
If n  2 , then T  F1 shown in the figure (1). In
  SR  T    1.
this case, n  6 , l  s  3 and  SR  T   7 .
Since diam  T   4 , we get n  3 . Then under
Therefore
 SR  T   7
the hypothesis
 SR  T    SR  T    1
30
6 n  l   s  
 1 4
4
6n  l  s
6  n  2    l  1   s  1   .
 1 4
4 Now, assume that n  3 . Therefore based on
6n  l  s
 6 n  l   s 
inductive hypothesis  SR  T   
.
4 .
Now, let f   xd  2   3 . So, the function
4
Any  SR  T    function can be extended to an
f  V0   xd  1 ,  , V2   xd  , V3  is an SRDF on tree T by assigning the weight 3 to x d  2 ,
SRDF on tree T . Thus by hypothesis we have
 SR  T   f  V 
2 to x d and 0 to x d  1 and u . Thus

 f V   2
 SR  T    SR  T    5 .
If deg  xd  3   2 , then l   l  1 and s   s  2
  SR  T    2
. Therefore
6 n  l   s   SR  T    SR  T    5
 2
4
6 n  l   s 
6  n  2    l  1   s  1   5
 2 4
6  n  4    l  1   s  2 
4
6n  l  s  5
 . 4
4
6n  l  s
Therefore, we can assume that for each  .
 SR  T    function 4
Now, let deg  xd  3   3 . In this case, l   l  2
, the weight of the vertex
x d  2 is equals to 0 .
and s   s  2 . Hence
Assume that x d  2 is a support vertex. Based on
Fact, T has only an adjacentleaf .we consider u as
an adjacent leaf to x d  2 . If x d  2 has a support
169
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
 SR  T    SR  T    5 u , v and x d  1 and 0 to all their neighboring
6 n  l   s  vertices. Therefore  SR  T    SR  T    9 .
 5
4 If deg  xd  3   2 , then l   l  2 and
6  n  4    l  2   s  2  s   s  3 . Hence
 5
4  SR  T    SR  T    9
6n  l  s
 . 6 n  l   s 
4  9
4
6 n  7    l  2    s  3 
Therefore in this case, if x d  2 is a support vertex,
6n  l  s  9
then  SR  T   . 4
4 6n  l  s
Now, assume that x d  2 is not a support vertex.  .
4
If x d  2 has three children u , v and w other than Now, let deg  xd  3   3 . In this case, l   l  3
x d  1 , then we put T   T   xd , xd  1  . We and s   s  3 . Thus
already assumed that for each  SR  T    SR  T    9
 SR  T    function the weight of x d  2 is
6 n  l   s 
equalsto 0 . We consider that the function f  is a  9
4
 SR  T    function , therefore f   xd  2   0 . 6 n  7    l  3    s  3 
Hence we can assume  9
4
f   u   f   v   f   w   3 and the child
6n  l  s
weight of each of the vertices u , v and w are 0 . In  .
4
this case, by changing the weight of the vertices u ,
Now, we support that x d  2 has only one support
v and w to 0 , childof each of the vertices u , v
and w to 3 and x d  2 to 3 , we obtain a child other than x d  1 . Let u be a support child
 SR  T    function where the weight of the x d  2 other than x d  1 . We put T   T  Txd  2 . So
vertex x d  2 is not equals to 0 which is a n  n  5 , l   l  2 and s   s  2 . Since
contradiction, since we previously assumed that for diam  T   4 , we get n  2 .
each  SR  T    function , we have the weight of If n  2 , then T  F3 shown in figure (1). So,
the vertex x d  2 is equals to 0. n  7 , l  s  3 and  SR  T   9 . Hence
So, we can assume that x d  2 has at most two 6n  l  s
 SR  T   .
support children other than x d  1 . 4
Now, let n  3 . So based on inductive hypothesis
First, assume x d  2 has two support children u and
6 n  l   s 
v other than x d  1 . We put T   T  Txd  2 . Since  SR  T    . Any
4
diam  T   4 , we get n  2 .  SR  T    function can be extended to an SRDF
If n  2 , then T  F2 shown in the figure (1). In on tree T by assigning the weight 3 to the vertices
this case, n  9 , l  s  4 and  SR  T   11 . u and x d  1 and 0 to all their neighboring vertices.
Thus  SR  T  
6n  l  s
.
So,  SR  T    SR  T    6 . And therefore
4
Now, let n  3 . In this case, n  n  7 and under
6 n  l   s 
the hypothesis  SR  T    . Any
4
 SR  T    function can be extended to an SRDF
on tree T by assigning the weight 3 to the vertices
170
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
 SR  T    SR  T    6 In the following, we characterize all the trees
6n  l  s
6 n  l   s  subjected to the condition  SR  T   . Let F
 6 4
4
be a family of trees T where it comes from a sequence of
6 n  5    l  2    s  2   
 6 trees T1 , T2 ,  , T j , j  1 such that T1  P4 or
4
T1  F3 (Shown in Fig 1) and if j  2 , then T j  1 can be
6n  l  s
 . obtained from T j with one of two operations O1 or O2 .
4

Case 2: deg x d 2   2. A. Operation O

1
u V T j   SR  T j  u    SR  T j 
T   T  Txd  2 . Since diam  T   4 , we
Let , and
We put
deg  u   2 . 𝑇In this case, T j  1 is obtained from T j by
get n  2 . 𝑗

If n  2 , then T  P5 . Thus adding a path P4 with the support vertex v and adding the
edge u v . (See Fig 2)
 SR  T   5
6n  l  s u
 .
4
Now, let n  3 . In this case, n  n  3 ,
l  l  1 and s  s  1 . Any
 SR  T    function can be extended to an SRDF 𝑣

on tree T by assigning the weight 3 to the vertices

u and x d  1 and 0 to all their neighboring vertices. Figure 2.Operation O1 .`
Therefore  SR  T    SR  T    3 . Hence with
the hypothesis we have B. Operation O 2
 SR  T    SR  T    3 Let u V T j  ,  SR  T j  u    SR  T j  and
6 n  l   s  deg  u   2 . In this case, T j  1 is obtained from T j by
 3 𝑇𝑗
4 adding a tree F3 with adding the edge u v where v is a
6 n  3    l  1    s  1 
 3 central vertex of tree F3 . (See Fig 3)
4
6n  l  s 𝑢
𝑣
 .
4
So the problem is solved. □

Figure 3.Operation O2 .

To prove that each tree T  F satisfy the condition

6n  l  s
F1 F2 F3  SR  T   , the following two Lemmas will be
4
F F
Figure 1.Trees F1 , F2 and F3 . F
useful. For each k  1 , let n k , l k and s k denote order,
1 1 1 number of leaves and number of support vertices of tree Tk ,
respectively.

171
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
6nj  lj  sj 6nj  lj  sj
Lemma 1: Let  SR  T j   and T j  1 is Lemma 2: Suppose that  SR  T j   and
4 4
obtained by T j with operation O1 , then T j  1 is obtained with the operation O2 of T j , then
6nj 1  l j 1  s j 1 6nj 1  l j 1  s j 1
 SR  T j  1   .  SR  T j  1   .
4 4
Proof: Proof:
Suppose that the path P4  xv y z and the vertex u Let F3 be a tree with central vertex v and vertex
is operation dependent. Any  SR  T j   function can be u  V  T j  is dependent to operation O2 . Suppose that f
extended to an SRDF on tree T by assigning the weight 3 to is a  SR  T j  1   function .
v , 2 to z and 0 to x and y . Therefore
f  u   0 , then f
 SR  T j  1    SR  T j   5 . Now, suppose that f is a
If is an SRDF on tree T j . Thus
 
Tj

 SR  T j  1   function .  SR  T j   f V Tj

If f  u   0 , then f is an SRDF on tree T j . So, in this  f  V   f  V  F3  .

f  u   0 . In this case, we have
Tj

 
case, we have Now, let
 SR  T j   f V  SR  T j    SR  T j  u 
 
Tj

 f  V   f  V  P4  .  f V Tj  u

Now, assume that f  u   0 . Hence f is an SRDF  f  V   f  V  F3  .

 
Tj  u

on T j  u . Therefore  SR  T j  u   f V and so On the other hand, always f  V  F3    9 .

Tj  u
So, in both cases we have
by the assumption
 SR  T j   f  V   f  V  F3 
 SR  T j    SR  T j  u 
 f V  Tj  u
  f V   9
  SR  T j  1   9 .
 f  V   f  V  P4  . Also any  SR  T j   function can be extended to an SRDF
On the other hand, always f  V  P4    5 .
on tree T by assigning the weight 3 to the support vertices of
So, in both cases we have
 SR  T j   f  V   f  V  P4  tree F3 and 0 to other vertices of tree F3 . Thus
 SR  T j  1    SR  T j   3 S  F3 
 f V   5
  SR  T j   9 .
  SR  T j  1   5 .
Therefore  SR  T j  1    SR  T j   5 . Since
Therefore  SR  T j  1    SR  T j   9 . Clearly
n j 1  n j  7 deg  u   2 , we get
deg  u   2 , we get l j  1  l j  2 and s j  1  s j  2 . and since

So, by induction we have l j  1  l j  3 and s j  1  s j  3 . So, by induction we

 SR  T j  1    SR  T j   5 have
 SR  T j  1    SR  T j   9
6nj  lj  sj
 5 6nj  lj  sj
4  9
6 n j 1  4   l j 1  2    s j 1  2  4
 5 6 n j 1  7   l j 1  3    s j 1  3 
4  9
6n j 1  l j 1  s j 1 4
 . 6n j 1  l j 1  s j 1
4  .
4
Now, hence the proof. □
Now, hence the proof. □
172
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
Theorem 7: For any tree T of order n  3 with l leaves Now, let T  F . In this case,
6n  l  s 6 n  l   s 
and s support vertices,  SR  T   if and only  SR  T    . Any
4 4
if T  F .  SR  T    function can be extended to an SRDF
by assigning the weight 3 to v , 0 to u and x d  1
Proof:
6n  l  s and 2 to x d . Thus
Suppose that  SR  T   . We proceed
 SR  T    SR  T    5
4
by an induction on the order n of a tree T . 6 n  l   s 
 5
If diam  T   3 , then based on proof of Theorem 6, we get 4
T  P4 and thus T  F . 6 n  4    l  2    s  2 
 5
Now, let diam  T   4 . We root the tree T at vertex x 0 . 4
Suppose that P  x0 x1  xd isa diagonal path. 6n  l  s
 .
Based on proof of Theorem 6, T does not have a Strong 4
support vertexand if T   F3 , then only in following two 6n  l  s
Therefore  SR  T   which is a
6n  l  s 4
cases are  SR  T   holds: contradiction.
4 Hence T   F . Thus T is obtained from T  with

Case 1: deg x d 2   3 , deg  x d 2   3 and x d 2

operation O1 .
is a support vertex.
Case 2: deg x d 2   4 , deg  x d 2   3 and x d 2

In this case, we put T   T  Txd  2 . Let u be has exactly two support children u and v other than
xd  1 .
adjacent leaf to x d  2 . Hence Txd  2  P4 . To prove
that T is obtained from T  with operation O1 , it is We put T   T  Txd  2 . In this case, Txd  2  F3 .
enough to show  SR  T   xd  3    SR  T   . Any  SR  T    function can be extended to an
Let T   F . On contrary, suppose that
3 to x d  1 , u and v
 SR  
SRDF by assigning the weight
T   xd  3   SR  T   . Any
0 to all their neighboring
 SR  T   xd  3   function
and vertices. Thus
can be extended to  SR  T    SR  T    9 .
an SRDF on tree T by assigning the weight 3 to 6 n  l   s 
x d  2 , 2 to x d and 0 to x d  1 , x d  3 and u . If T  
 F , then  SR  T    .
4
Therefore  SR  T    SR  T   xd  3   5 . Thus Therefore
 SR  T    SR  T    9
 SR  T    SR  T   x d  3   5
6 n  l   s 
  SR  T    5  9
4
6 n  l   s  6 n  7    l  3    s  3 
 5  9
4 4
6 n  4    l  2    s  2  6n  l  s
 5  .
4 4
6n  l  s 6n  l  s

4
. So,  SR  T   which is a
4
6n  l  s
Therefore  SR  T   which is a contradiction. Hence T   F .
4 To prove that T is obtained from T  with operation
contradiction. So,  SR  T   xd  3    SR  T   . O2 , it is enough to show that

173
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 8 164 – 174
_______________________________________________________________________________________________
 SR  T   xd  3    SR  T   . Any Also, if T is obtained from T  with operationO2 , then from
 SR  T    function can be extended to an SRDF the Lemma 2 it follows that  SR  T  
6n  l  s
.
by assigning the weight 3 to x d  2 , 0 to x d  3 and 4
vertices 
S Tx d  2  and 
2 to L Txd  2 . Thus
Hence the proof. □

 SR  T    SR  T   xd  3   9 . Therefore REFERENCES

 SR  T    SR  T   x d  3   9
[1] Cockayne, E. J., Dreyer, P. A., Hedetniemi, S. M., and
Hedetniemi, S. T. (2004). Roman domination in graphs. Discrete
  SR  T    9 Mathematics, 278(1):11-22.
[2] De Jaenisch, C.F. (1862) Applications de l’Analyse
6 n  l   s 
 9 Mathematique au Jeu des Echecs. Petrograd.
4 [3] Dreyer, P. A. (2000). Applications and variations of domination
6 n  7    l  3    s  3  in graphs. The State University of New Jersey, New Brunswick,
 9 NJ.
4 [4] Haynes, T. W., Hedetniemi, S. T., and Slater, P. J. (1998).
6n  l  s
 . Fundamentals of domination in graphs, volume 208
4 ofmonographs andtextbooks in pu
6n  l  s -reand applied mathematics.
Thus  SR  T   which is a [5] Haynes, T. W., Hedetniemi, S. T., and Slater, P. J. (1998).
4 Domination in graphs: advanced topics, volume 209 of
contradiction.
 xd  3    SR  T   . Therefore T is
monographs and textbooks in pure and applied mathematics.
So  SR T   [6] Henning, M. A. (2003). Defending the roman empire from
obtained from T  with operation O2 . Hence T  F multiple attacks. Discrete Mathematics, 271(1):101-115.
. [7] Pavlic, P. and Zerovnik, J. (2012). Roman domination number of
Hence in both cases T  F . the cartesian products of paths and cycles. The electronic journal
of combinatorics, 19(3):P19.
Conversely, let T  F . We apply induction on the
[8] Selvakumar, K. and Kamaraj, M. (2016). Strong roman
number of operations performed to construct a tree T . domination in graphs. International Journal of Mathematical
6n  l  s
T  P4 or T  F3 , then clearly  SR  T  
Archive (IJMA) ISSN 2229-5046, 7(3).
If .
4 [9] Singh, G. S. (2010). Graph Theory. PHI Learning Private
Now, let T  P4 and T  F3 . Based on the structure of F , Limited, New Delhi.
[10] Stewart, I. (1999). Defend the roman empire! Scientic American,
let T be obtained of T   F with operations O1 and O2 . 281:136-138.
6 n  l   s 
Under the hypothesis we have  SR  T   
[11] Sumenjak, T. K., Pavlic, P., and Tepeh, A. (2012). On the roman
4 domination in the lexicographic product of graphs. Discrete
  
where n , l and s denote order , number of leaves and
Applied Mathematics, 160(13):2030-2036.

number of vertices of tree T  , respectively. [12] Xueliang, F., Yuansheng, Y., and Baoqi, J. (2009). Roman
domination in regular graphs. Discrete Mathematics,
If T is obtained from T  with operation O1 , then based on 309(6):1528-1537.
6n  l  s
Lemma 1 we have  SR  T   .
4

174
IJRITCC | August 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________