0% found this document useful (0 votes)
192 views9 pages

Bounds On The Distance Two-Domination Number

This document presents bounds on the distance two-domination number (c2) of a graph. It begins with definitions of distance two-domination and c2. Several theorems are then presented that establish upper bounds for c2 in terms of the order (number of vertices) of the graph. Specifically, it is shown that c2 is at most 1, at most 2/5 the order for certain graphs, and at most the floor of the order divided by 3 for graphs where each component has at least 3 vertices. A characterization is given for when c2 equals the order divided by 3.

Uploaded by

vpro1
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
192 views9 pages

Bounds On The Distance Two-Domination Number

This document presents bounds on the distance two-domination number (c2) of a graph. It begins with definitions of distance two-domination and c2. Several theorems are then presented that establish upper bounds for c2 in terms of the order (number of vertices) of the graph. Specifically, it is shown that c2 is at most 1, at most 2/5 the order for certain graphs, and at most the floor of the order divided by 3 for graphs where each component has at least 3 vertices. A characterization is given for when c2 equals the order divided by 3.

Uploaded by

vpro1
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

Graphs and Combinatorics (2002) 18:667675

Graphs and
Combinatorics
Springer-Verlag 2002

Bounds on the Distance Two-Domination Number


of a Graph

N. Sridharan1 , V.S.A. Subramanian2 , and M.D. Elias3


1
Department of Mathematics, Alagappa University, Karaikudi-630 003, Tamilnadu, India
2
A.P.S.A. College, Tiruppatur, India
3
Department of Mathematics, B.U.E.T., Dhaka, Bangladesh

Abstract. For a graph G = (V , E), a subset D  V G is said to be distance two-domi-


nating set in G if for each vertex u 2 V  D, there exists a vertex v 2 D such that du; v  2.
The minimum cardinality of a distance two-dominating set in G is called a distance two-
domination number and is denoted by c2 G. In this note we obtain various upper bounds
for c2 G and characterize the classes of graphs attaining these bounds.

1. Introduction

We consider only nite undirected graphs without loops or multiple edges. The
order of graph G V ; E is jV Gj and its size is jEGj. If S  V G then hSi
denotes the induced subgraph of G with vertex set S. Two vertices in S are
adjacent in hSi if and only if they are adjacent in G. A vertex with degree one is
called an end-vertex or a pendant vertex. The minimum and maximum degrees of
vertices of G are denoted by d and D respectively. For every vertex v 2 V G the
neighbourhood N1 v is the set of all vertices adjacent to v in G; i.e.,

N1 v fu 2 V G : uv 2 EGg:

The distance du; v between two vertices u and v is the length of a shortest u-v
path in G. If there is no u-v path in G, then du; v 1. For every vertex
v 2 V G, the kth neighbourhood Nk v of v is dened by

Nk v fu 2 V G : du; v k in Gg:

For terminology and notation not given here, the reader is referred to [2] or [5].
The concept of domination was generalized to distance k-domination. Slater
[9] termed a distance k-domination set as a k-basis.
Denition (Slater): Given a nite graph G V ; E, a subset D of the set V of
vertices is called a dominating set when every vertex not in D is adjacent to atleast
668 N. Sridharan et al.

one vertex in D. More generally, a subset B of V is called a k-basis (k  1) if for


each vertex v of V there is at least one vertex b in B such that the distance between
b and v in G, denoted by dG b; v, is less than or equal to k. Thus a dominating set
is a 1-basis.
Slater [9] also gave an interpretation for a k-basis in terms of communication
network. For more information on domination, total domination and distance
k-domination in graphs the reader is referred to [1, 3, 4, 6, 7, 8, 9].
In this note, we nd upper bounds for distance two-domination number for
connected graphs.

Denition. Let G be a graph. A subset D of V G is said to be a distance two-


dominating set for G if for each u 2 V  D; du; v  2, for some v in D. A minimum
cardinality of a distance two-dominating set is said to be distance two-domination
number of G and is denoted by c2 G. A distance two-dominating set D of G is said
to be a c2 set if jDj c2 G.

Examples.
1. c2 Kn 1
2. c2 Km;n
 1
3. c2 Pn n5  where Pn is the path on n vertices
4. c2 Cn n5 where Cn is a cycle on n vertices
5. If G is the Petersons graph, c2 G 2
We rst observe that c2 G  cG and c2 G  cG2 where V G2 V G and
u; v 2 V are adjacent in G2 if and only if du; v  2 in G.

2. Results

 1, where G
Theorem 1. For any graph G, either c2 G 1 or c2 G  is the com-
plement graph of G.

Proof. Assume that c2 G 6 1. Choose vertices u and v in V G such that


,
du; v > 2 in G. So in G

(i) u is adjacent to every w where w 2 = NG u [ fug and


 for all w 2 NG u, du; w 2 in G
(ii) u v w is a path in G . Hence fug is a c2 set
.
in G
It is known [8] that if G is a connected graph of order p with minimum degree 2
and if G is not isomorphic to any one of the seven graphs in Figure 1, then
c  2=5p.
If G is a connected graph of order p with minimum degree 2 and if G
is not isomorphic to any of the seven graphs in Figure 1, then
c2 G  cG  2=5p. Also for these seven graphs c2 G  2=5p. Thus we have
the following theorem.
Bounds on the Distance Two-Domination Number of a Graph 669

Fig. 1.

Theorem 2. If G is connected with order p and with minimum degree 2, then


c2 G  2=5p.
In the following theorem we obtain a stronger result.

Theorem 3. If each component of G contains at least three vertices then


c2 G  b p=3c, where p is the order of G.

Proof. We rst prove this result for a tree. The result is clearly true for a tree
whose order is 3 or 4 or 5. We assume that the result is true for all trees of order
 n. Let T be a tree with n+1 vertices. Let P = u0 ; u1 ; u2 ; . . . ; um be the longest
path in T . If m  2 then c2 T 1  bn 1=3c. So assume that m  3. Let T 0 be
the component of T  fum ; um1 ; um2 g which contains the vertex u0 . If the order
of T 0  2 then fum2 g is a c2 -set for T . If the order of T 0  3, by induction
c2 T 0  bn  2=3c. Let D0 be a c2 -set for T 0 . Then D0 [ fum2 g is a distance two-
dominating set for T and hence c2 T  bn 1=3c. Hence by induction the
result is true for all trees. Let G be any graph in which each component contains at
least three vertices. For each component H of G, c2 H  c2 T  P bOH =3c,
whenever
P T is a spanning tree in H . Hence c 2 G c2 H 
bOH =3c  b p=3c.

Now we characterise the graphs for which c2 p=3. For G = C6 , the cycle of
order 6, c2 G 2 p=3.

Denition 1. For a graph H, we denote by H the graph obtained from H by adding


a vertex v0 for each vertex v in H and joining v and v0 .
670 N. Sridharan et al.

Fig. 2. Graphs H, H+ and H+2

For every connected graph H , we dene a graph denoted by H 2 as


the graph obtained from H by attaching a free path of length two at every vertex
of H .
B.D. Acharya, H.B. Walikar and E. Sampathkumar [1] have proved the fol-
lowing theorem.

Theorem 4. Let G be a connected graph of order p  2. Then c p=2 i G C4 or


H for some connected graph H.
We obtain a similar result for c2 .
We note that if H is a connected graph of order p, then c2 H 2 p=3. We
claim that these are the only graphs other than C3 ; C6 for which c2 G p=3.

Theorem 5. For a connected graph G of order p, c2 G p=3 if and only if either


G C3 ; C6 or G H 2 for some connected graph H.

Proof. Let G be a graph with c2 G p=3. The result is true if G = P3 , the path
on three vertices. So assume that G 6 C3 ; P3 . Let A be a distance two-dominating
set with jAj c2 G p=3 and with the minimum number of components (i.e.,
number of components in hAi is minimum). Then we note the following.
(i) jAj p=3 and so N1 A 6 /
(ii) If a vertex x in A is not adjacent to any vertex of N1 A, then there is
atleast one vertex y in A such that xy is an edge in G and hence A  fxg is also a
distance two-dominating set with less than p=3 vertices which is a contradiction.
Therefore every vertex in A is adjacent to atleast one vertex in N1 A
(iii) jN1 Aj  p=3. Otherwise by (ii) it follows that N1 A is a distance two-
dominating set with cardinality less than c2 G which is a contradiction.
(iv) We claim that N2 A 6 /. If possible assume that N2 A /. Then every
vertex in A is isolated in hAi [if a vertex x in A is not isolated in hAi, then A  fxg is
a distance two-dominating set with p=3  1 vertices as N2 A / which will be a
contradiction]. If there is one y in N1 A which is adjacent to two distinct vertices
x1 and x2 of A, then A [ fyg  fx1 ; x2 g is a distance two-dominating set, which is
a contradiction.
Bounds on the Distance Two-Domination Number of a Graph 671

Therefore no vertex in N1 A is adjacent to two distinct vertices of A. To each


x 2 A, let Sx fy 2 N1 A=xy is an edge in Gg. i.e., Sx = N1 x \ N1 A. Fix
one x1 2 A. As G 6 C3 ; P3 , we have x1 [ Sx1 6 G and as G is connected, there is
some y in Sx1 and z in Sx2 for some x2 ; x2 6 x1 in A such that yz is an edge in G.
Then A  fx1 ; x2 g [ fy; zg is a c2 -set with fewer components, which is again a
contradiction. Thus N2 A 6 /.
(v) Let S1  N1 A be a set with least cardinality such that every vertex of
N2 A is adjacent to atleast one vertex of S1 . Let S2  A be a set with least
cardinality such that every vertex of S1 is adjacent to atleast one vertex of S2 .
Let S1 fy 2 N1 A=xy 2 EG for some x in S2 g then jS2 j  jS1 j  jS1 j and
jS2 j  jS1 j  jN2 Aj we note that jS2 j  jS2 j jS1 j jN2 Aj=3. If x1 2 A  S2
such that x1 is adjacent to some vertex x2 of A, then A  fx1 g itself is a distance
two-dominating set with p=3  1 elements, which is a contradiction. Thus
x 2 A  S2 ) x is isolated in hAi.
(vi) There is no y in N1 A such that y is adjacent to two distinct vertices of
A  S2 [For, if y in N1 A is adjacent to x1 ; x2 in A  S2 , then A  fx1 ; x2 g [ fyg is
a distance two-dominating set, leading to contradiction]. If x1 ; x2 are distinct
vertices in A  S2 and y1 ; y2 2 N1 A such that x1 y1 , x2 y2 are edges in G, then y1 y2
is not an edge, for if y1 y2 were an edge then A  fx1 ; x2 g [ fy1 ; y2 g is a distance
two-dominating set with less number of connected components. Let x 2 A,
y 2 N1 A  S1 , deg(x) = 1 in G and xy is an edge in G. Then y is isolated in
hN1 Ai. [For if there is one y 0 2 N1 A such that yy 0 is an edge in G, then y 0 2 S1 .
Thus A  fxg [ fy 0 g is a distance two-dominating set with fewer components,
which is a contradiction].
(vii) Now consider the case in which there is a vertex x 2 A  S2 with deg(x)
=1 in G. Let U1 fa 2 A  S2 /deg(a) = 1g. Select one a1 2 U1 . Select a path
a1 y1 z1 where y1 2 N1 A (In fact y1 2 N1 A  S1 ) and z1 2 N2 A. Select a path
a0 y0 z1 where a0 2 S2 .
Let V0 fa 2 U1 / there is some z 2 N2 A such that da; z 2 and
dz; a0 2g. To each a 2 V0 select one V0  N2 A path a y z such that
dz; a0 2. Thus we have selected jV0 j paths. These jV0 j paths are distinct. [For if
a1 6 a2 2 V0 and a1 y1 z1 and a2 y2 z2 are selected paths and if z1 z2 then
A  fa1 ; a2 g [ fzg is a distance two-dominating set with less than p/3 vertices,
which is a contradiction].
Let V2 fz=z 2 N2 A and z is in any one of these pathsg. Note that
A  fa0 g [ V0 [ V2 is not a distance two-dominating set (as it contains less
than p/3 vertices). So there exists a path a0 y2 z0 where y2 2 N1 A and z0 2 N2 A
such that dz; z0  3 8 z 2 V2 .
Let Va1 V0 [ fall points on the selected jV0 j pathsg [ fa0 ; y0 ; z0 ; y2 g. Note that
jVa1 \ Aj < 1=3jVa1 j. Select a2 2 U1  V0 . Form Va2 as above. Proceeding like this,
we get Va1 ; Va2 ; . . . ; Vam such that they are disjoint and
jVai \ Aj < 1=3jVai j; 8 i 1; 2; . . . ; m and
[m
U1  Vai :
i1
672 N. Sridharan et al.

To each a 2 A  S2 [ U1 , we have a 2 A  S2 and deg(a)  2. To each such a


we associate Wa where Wa fag [ N a. Clearly Wa \ A = fag and jWa j  3. Now
to each aj 2 S2 , by minimality of S2 there exists yj in S1 such that N yj \ S2 =
faj g and again by minimality of S1 there exists zj 2 N2 A such that N zj \ S1 =
fyj g. Select one such yj and zj . Put Xaj faj ; yj ; zj g. Then Xaj \ A = faj g and jXaj j
= 3. Now V = Q [ R [ T where

[
m
Q Vai ;
i1

S S
R fW a : a 2 A  S2 [ U1 g and T fXaj : aj 2 S2  Qg is a partition of
V V G. Note that jAj = jV \ Aj < 1=3jV j which is a contradiction. Thus
a 2 A  S2 ) deg(a) 6 1 in G.
We have already noted that

jS2 j jS1 j jN2 Aj


jS2 j  :
3

jS2 jjS1 jjN2 Aj


If jS2 j < 3 ; there should be atleast one vertex a 2 A  S2 with deg(a) =
1. So,

jS2 j jS1 j jN2 Aj


jS2 j
3

This is possible only when jS2 j jS1 j jS1 j jN2 Aj. To each ai 2 S2 , select
one yi 2 S1 and zi 2 N2 A such that N zi \ S1 fyi g and N yi \ S2 = fai g. Let
Xai fai ; yi ; zi g 8 ai 2 S2 . If ai 6 aj then Xai \ Xaj /. Now we claim that A =
S2 . If possible assume that S2  A. To each a 2 A  S2 let Wa fag [ N a. For
a; b 2 A  S2 and a 6 b, Wa \ Wb / (otherwise A  fa; bg [ fug is a distance
two-domination set where u 2 Wa \ Wb ). Also for each a in A  S2 , N a
 N1 A.
Now
! !
[ [ [
V G Xa Wa ;
a2S2 a2AS2

otherwise jAj < p=3. As G is connected, we can nd one ai 2 S2 and a 2 A  S2


such that there is an edge joining Xai and Wa . Let xy be an edge with x 2 Xai
and y 2 Wa . Clearly y 2 N1 A  S1 N1 A  S1 , so x 2
= A. Hence either x = yi
or x = zi . If x = yi then A  fai ; ag [ fyg is a distance two-dominating set
with less than p/3 vertices which is a contradiction. Hence x = zi . If ai is
Bounds on the Distance Two-Domination Number of a Graph 673

isolated in hAi then A  fai ; ag [ fzi ; yg is a c2 -dominating set with fewer


components. Hence ai is not isolated in hAi. Hence A  fai ; ag [ fyg is a
distance two-dominating set with less than p/3 vertices which is also a con-
tradiction.
[ A  S2 /: Thus S2 A and N1 A S1 :
[ jAj = jS2 j jS1 j jN2 Aj p=3 and
[
V Xa :
a2A

It follows that every vertex in N1 A is isolated in hN1 Ai and if yi 2 Xai then


N1 yi \ A = fai g and N1 yi \ N2 A = fzi g. If p = 6, as G is connected, G is
either C6 or a path of length 5. Assume that p > 6 then we claim that every
vertex in N2 A is isolated in hN2 Ai. If possible let zi zj be an edge in
N2 A. If both ai and aj are isolated in A, then A  fai ; aj g [ fzi ; zj g will be a
distance two-dominating set with less number of components leading to a
contradiction. If there is one ak k 6 i; j adjacent to atleast one of ai and aj say
ak ai is an edge in G, then A  fai ; aj g [ fzj g is a distance two-dominating
set with less than p/3 points which is a contradiction. So there is no
ak k 6 i; j adjacent to atleast one of ai ; aj . Also we have ai ; aj are not isolated
in A.
So ai aj is an edge and there is some zk k 6 i; j adjacent to atleast one of zi and
zj say, zk zi is an edge in hN2 Ai. Then A  fai ; ak g [ fzk g is a distance two-
dominating set with less than p/3 points which is a contradiction.
Therefore every vertex of N2 A is isolated in hN2 Ai. As G is connected, the
only possibility is that hAi is connected. Put H = hAi. Then H is connected and
G = H 2 .

Theorem 6. If G has p vertices and no isolated vertex, then c2  p  D 1=2.

Proof. Let u be a vertex with du D and X V G  fug [ N1 u [ N2 u.


Let S be the set of isolates of hX i. Then each y in S is adjacent to atleast
onevertex in N2 u. Let MS  N2 u such that MS is the set with
minimum cardinality such that each vertex of S is adjacent to atleast one vertex
of MS and M2 S  N1 u with minimum cardinality such that each vertex of
MS is adjacent to atleast one vertex of M2 S. Therefore jM2 Sj 
jMSj  jSj.
Let A be a distance two-dominating set for hX  Si. Then jAj jX  Sj=2 if
hX  Si = mK2 for some m otherwise jAj < jX  Sj=2. Now fug [ M2 S [ A is a
distance two-dominating set for G.
Therefore

c2 G  1 jM2 Sj jAj
c2 G  1 jM2 Sj jX  Sj=2
674 N. Sridharan et al.

jX  Sj 2jM2 Sj jX j  jSj 2jM2 Sj


c2 G  1 1
2 2
jX j jM2 Sj
1 as jM2 Sj  jMSj  jSj
2
p  D 1
1 as X and MS  V  fug [ N1 u
2
pD12

2
pD1
 :
2

Remarks. (1) For G = K1;p1 , c2 G = 1 = p  D 1=2 as D p  1


(2) Assume that D 1 < p (i.e. N2 u 6 /). If S / and X 6 / then each
component of hX i is adjacent to atleast one vertex of N2 u. Let M be a subset of
N2 u with least cardinality such that each component of hX i is adjacent to atleast
one vertex of M.
Then c2 hX [ Mi  jX j jMj=3  p  D 1=3. As fug [ fa distance two-
dominating set of hX [ Mig is a distance two-dominating set for G. We get

c2 G  1 p  D 1=3 p  D 2=3
i.e., c2 G  p  D 2=3 < p  D 1=2 as D 1 < p:

D 1 < p ) 3D  2D 4  3 < 3p  2p
) 2p  2D 4 < 3p  3D 3 ) p  D 2=3 < p  D 1=2

(3) If X / then c2 G 1  p  D=2 as N2 u 6 /. So c2 G <


p  D 1=2
(4) If S 6 / then each component of hX  Si is adjacent to atleast one vertex
of N2 u. Let M be a subset of N2 u with least cardinality such that each com-
ponent of hX  Si is adjacent to atleast one vertex of M.
Let S1 = fx 2 S = x is adjacent to atleast one vertex of Mg.
Let S2 S  S1 :
Let M1 S2 be the least cardinality set in N2 u such that each vertex of S2
is adjacent to atleast one vertex of M1 S2 . Note that M1 S2 \ M /. Let
M2 S2  N1 u with least cardinality such that each vertex of M1 S2 is adja-
cent to a vertex of M2 S2 . Note that jM2 S2 j  jM1 S2 j  jS2 j and
jM2 S2 j  D. Let l = jM [ hX  S2 ij. Let A be a c2 set for hM [ X  S2 i.
Then jAj  l=3.

Case (i). Let l  6. Then fug [ M2 S2 [ A is a distance two-dominating set for


G.
Bounds on the Distance Two-Domination Number of a Graph 675

Therefore
c2 G  1 jM2 S2 j jAj
 1 jM2 S2 j l=3
 1 l=3 p  D  1  l=2 jM2 S2 j  M1 S2 S2 =2
 p  D  1=2 1  l=2 l=3
 p  D  1=2 6  l=6  p  D  1=2 as l  6

Case (ii). Let l  5 then jAj = 1 and

c2 G  1 jM2 S2 j 1
 2 jM2 S2 j
 2 jM1 S2 j jS2 j=2
 2 p  D  1  l=2
 p  D  1 4  l=2
 p  D=2 as 3  l  5

Thus we get the following theorem.


Theorem 7. If G has p vertices (p  3) and no isolated vertex, then
c2 G p  D 1=2 if and only if D p  1 and
c2 G < p  D 1=2 if D 6 p  1:

References

1. Acharaya, B.D., Walikar, H.B., Sampathkumar, E.: Recent development in the theory
of domination in graphs. Metha Research Institute, Allahabad MRI Lect. Notes Math.
1979
2. Bondy, J.A., Murty, U.S.R.: Graph Theory and Applications. New York: North Holland
1986
3. Cockyane, E.J., Hedetniemi, S.T.: Towards a theory of dominations in graphs. Net-
works 7, 243261 (1977)
4. Cockyane, E.J., Dawes, R.M., Hedetniemi, S.T.: Total domination in graphs. Networks
10, 211219 (1980)
5. Harary, F.: Graph Theory. Reading: Addison-Wesley 1969
6. Haynes, T.W., Hedetniemi, S.T., Slater, P.J.: Fundamentals of Domination in Graphs.
New York: Marcel Dekkar 1998
7. Haynes, T.W., Hedetniemi, S.T., Slater, P.J. (eds.): Domination in Graphs: Advanced
Topics. New York: Marcel Dekkar 1998
8. Mc Cuaig, W., Shepherd, B.: Domination in graphs with minimum degree two.
J. Graph Theory 13, 749762 (1989)
9. Slater, P.J.: R-dominations in graphs. J. Assoc. Comput. Macg. 23, 446460 (1976)

Received: May 31, 1999


Final version received: July 13, 2000

You might also like