ON THE LOCATING-CHROMATIC NUMBERS OF NON-HOMOGENEOUS CATERPILLARS AND FIRECRACKER GRAPHS
Far East Journal of Mathematical Sciences (FJMS) © 2016 Pushpa Publishing House, Allahabad, India Published Online: September 2016 http://dx.doi.org/10.17654/MS100081305 Volume 100, Number 8, 2016, Pages 1305-1316
ISSN: 0972-0871
ON THE LOCATING-CHROMATIC NUMBERS OF
NON-HOMOGENEOUS CATERPILLARS
AND FIRECRACKER GRAPHS
Asmiati Department of Mathematics Faculty of Mathematics and Natural Sciences Lampung University Jl. Brojonegoro No. 1, Gedung Meneng Bandar Lampung Indonesia. e-mail: asmiati308@yahoo.com
Abstract
We determine the locating-chromatic numbers of non-homogeneous
caterpillars and firecracker graphs.1. Introduction
The notion of locating-chromatic number of a graph was introduced by
Chartrand et al. [8]. Let G be a finite, simple, and connected graph. Let c be a
C C CV G
Π =
proper k-coloring of G and { , , ..., k } be a partition of ( )
1
2 V G C
induced by c on ( ) , where i is the set of vertices receiving color i. The
color code c v d v C d v C d v C ( ) of v is the ordered k-tuple ( ( , ) ( , , ) , ..., ( , k ) ) ,
Π
1
2 d v C d v x x C
where ( , ) = min { ( , ) | ∈ } for any i. If all distinct vertices of G i i
Received: May 13, 2016; Accepted: June 29, 2016 2010 Mathematics Subject Classification: 05C12, 05C15. Keywords and phrases: locating-chromatic number, caterpillar, firecracker graph.
1306 Asmiati have distinct color codes, then c is called a locating-chromatic k-coloring of
G (k-locating coloring, in short). The locating-chromatic number, χ ( ) G is
L the smallest k such that G has a locating k-coloring.Chartrand et al. [8] determined the locating-chromatic number for paths, cycles, complete multipartite graphs and double stars. Behtoei and Omoomi [7] discussed the locating-chromatic number for Kneser Graph. Specially for amalgamation of stars, Asmiati et al. [1, 4] determined locating-chromatic number for homogeneous amalgamation of stars and non-homogeneous amalgamation of stars, respectively. Furthermore, the locating-chromatic number of the operation of two graphs is discussed by Baskoro and Purwasih [6]. They determined the locating-chromatic number for a corona product of two graphs.
Chartrand et al. [8] characterized graphs have locating-chromatic number −
n
1 . They also determined graphs whose locating-chromatic numbers are bounded by n − 2 . Moreover, Asmiati and Baskoro [3] characterized all maximal graphs containing cycle. In general, characterization of all tress with locating-chromatic number 3 is given in Baskoro and Asmiati [5].
Asmiati et al. [2] determined the locating-chromatic number for homogeneous firecracker graphs, Motivated by these results, we determine the locating-chromatic numbers of non-homogeneous caterpillars and firecracker graphs.
The following results were proved by Chartrand et al. in [8]. We denote the set of neighbors of a vertex v in G by N ( )
V .
Theorem 1 [8] . Let c be a locating-coloring in a connected graph G
=
(
V , E ) . If u and v are distinct vertices of G such that d ( u , w ) d ( v , w ) for
=
all w ∈ V ( ) { G − u , v } , then c ( ) u ≠ c ( ) v . In particular, if u and v are non- adjacent vertices of G such that N ( ) u = N ( ) v , then c ( ) u ≠ c ( ) v .
Corollary 1 [8] . If G is a connected graph containing a vertex adjacent to m end -vertices of G, then χ ( ) G ≥ + m 1 .
L
On the Locating-chromatic Numbers of Non-homogeneous … 1307 Corollary 1 gives a lower bound for the locating-chromatic numbers of a general graph G.
2. Locating-chromatic Number of Non-homogeneous Caterpillar
In this section, we discuss about the locating-chromatic number of non- homogeneous caterpillar. Let P be a path with
V ( ) { P = x , x , ..., x } m m
1 2 m
= and E ( ) { P x x , x x , ..., x x } . A non-homogeneous caterpillar is
m
1
2
2 3 m − 1 m =
obtained by connecting n pendant vertices ( a , j
1 , 2 , ..., n ) to one i ij i
≤ ≤ particular vertex x of path P , where 1 i m , which we denote by i m
C ( m ; n , n , ..., n ) . The non-homogeneous caterpillar C ( m ; n , n , ..., n )
1 2 m
1 2 m consists of vertex set V ( C ( m ; n , n , ..., n ) ) { = x | 1 ≤ i ≤ m } { a | 1 ≤ i
1 2 m i ∪ ij
, 1 ; , , ...,
1 ≤ m ≤ j ≤ n } and edge set E ( C ( m n n n ) ) { = x x | ≤ i ≤ m
1 | ≤ ≤ ≤ ≤ 1 } { x a
1 2 m i i
+ i
1 i m , 1 j n } . − ∪ i ij i
Let K , with the vertex x as the center, be a subgraph of 1 , n i i
C ( m ; n , n , ..., n ) . We denote the set of vertices and edges by V ( K ) =
1 2 m 1 , n i
|1 ≤ ≤ { a j n } { } x and E ( K ) { = x a |
1 ≤ j ≤ n } , respectively. Thus,
ij i ∪ i 1 , n i ij i i
C ( m ; n , n , ..., n ) contains m stars K with x as a center. If n =
1 2 m 1 , n i max i
max { n , n , ..., n } , then subgraph K is called the maximum star
1 2 m 1 n , max
subgraph in the non-homogeneous caterpillar C ( m ; n , n , ..., n ) . If there
1 2 m
are p subgraphs K , then every subgraph, from left to right, are denoted
1 n , max i
≤ ≤ by K , where 1 i p .
1 , n maks
Definition 1. Let , ⊂ ; , , ..., , where 1 ≤ ≠ K K C ( m n n n ) i j
1 , n 1 , n
1 2 m i j
≤ = ≠ m . If n n n , such that i j max
(1) d ( x , x ) = d ( x , x ) , with x is the center of K , or i m j m m 1 n , max
- > +
- ≤ + = χ .
- ≤ n p Since the number of leaves in a maximal subgraph is ,
- ≥ χ
- ≤ n
- n Consider the
- n -coloring c on ( )
m C n n n m ..., , , ;
max
1
max
1
p .
for ≤
1
2
Next, we determine the upper bound of ( )
m C n n n m ..., , , ;
p
max
1
C n n n n m L m for .
1
2
max
by Corollary 1, ( ) ( ) , , 1 ..., , ;
max n
2
as follows: a. Find the number of subgraphs
1
1 are colored by , ..., , 3 , 2 , 1 p respectively.
1
} ( ) { } . \
..., , 3 , 2 , 1 n
max
1 are colored by {
K where p i ≤ ≤
1 i n
max ,
c. Leaves in ,
K x ∈ where p i ≤ ≤
max , 1 n
1 i n i
max ,
b. Vertices ,
1 p i ≤ ≤ respectively.
K where ,
1 i n
max ,
of the subgraphs from left to right as ,
K and denote it by p. Denote each
max
1
1 for .
1
( ) m
, 1 n K be the maximum star subgraph of
Theorem 2. Let max
are called star subgraphs with the same distance.
1
K ,
j n
and
K ,
2
i n
then subgraphs
K
max , 1 n
,
≠ , are the centers of
= with
x
and p p x x x, , , p j i x x d x x d
Asmiati 1308 (2) ( ) ( )
C n n n m ..., , , ;
1 and p be the number of subgraphs . max
2
1 ,
C n n n m ..., , , ;
( )
mFirst we determine the trivial lower bound of
C n n n m L m Proof.
1 n p if n n p if n
2
max max max max
1 ..., , , ;
2 , 1 ,
⎩ ⎨ ⎧
, 1 n K Then for
( ) ( )
is
1
2
C n n n m ..., , , ;
( ) m
max ≥ n the locating-chromatic number of non-homogeneous caterpillar
2
,
x c
- k
- k n
- n
- n
2
1
2
..., , , ;
m C n n n m
We show that the color codes for all vertices in ( )
j n
2
jl i il } . ..., ,
, 1 ..., 2 , = 1 = | ≠ | l a c n l a c
j i = x c x c then ( ) { } { ( ) ,
versa, if ( ) ( ) ,
j n then i x and j x should be given different colors. Vice
} , ..., ,
1
= | = = | l a c n l a c jl i il
1
maximum star subgraph and ( ) { } { ( ) , , 1 ..., 2 ,
j i
= n n have the same distance from thewith
1
K ,
j n
and
1
K ,
i n
i. If
for ,
max
is colored by colors that correspond with ,
i
x and . j xp C and . q
≠ because they have different distances from
( ) v c Π
A then ( ) u c Π
p A and , q
lie in different intervals, say
1
K ,
j n
and
1
K ,
i n
≠ because their color codes differ in the ordinate of colors
p are different. Let u , v , v u ≠
Π Π
then ( ) ( ) v c u c
n n n j i
max
1
and ( ) ( ) . v c u c =
K V v ∈
1 j n
K V ( ) , ,
1 i n
,
u ( ) ,
be the leaves, where ∈
1 − k T respectively.
1
C
1 max
1
from
max n
e. Let T = {all combinations ( )
K
1 p n
,
A be an open interval after . max
p k and 1 + p
1 − ≤ ≤
1
K where ,
1 ,
color}, such that T =
and ,
1
,
i n K max
between
i n A K be an open interval
max
,
1
1 ,
1 A be an open interval before
d. Let
On the Locating-chromatic Numbers of Non-homogeneous … 1309
max
{ }
K ,
1 A or ,
i n
1 3 + ≤ ≤ p k then every vertex of
k A where ,
lies in interval ,
1
K ,
i n
h. If
1 T respectively.
is colored by colors that are associated with ,
1
K ,
2
A then every vertex of i nlies in interval
1
1
K ,
i n
g. If
1 in the interval as defined in item d.
K ,
i n
f. Identify subgraph
∈ are color combination not containing the color i.
T T T with T T i
..., , ,
1 max
2
- ≤ n
- If ,
- = =
- If
- If
- If one of
- If ( )
- ≤ n p are different, thus c is a locating-
- ≤ χ n n n n m C
- ≤ n p
Figure 1. A minimum locating-coloring of C(9; 2, 3, 1, 1, 3, 1, 2, 3, 3).
- By Corollary 1, we have that ( ) ( ) .
- ≥ χ n n n n m C
- n colors are not enough. For a contradiction, assume that there exists a
- n -locating coloring c on
- > n
- > n
max
, 1 ..., , ;
1
max > n p .
for
1
2
m C n n n m ..., , , ;
max
Next, we show the lower bound of ( )
1
L m for .
1
2
max
coloring. So, ( ) ( ) , , 1 ..., , ;
max
1
2
1
L m
max
max max n h a c n h a c
jh ih
1
such that ( ) { } { ( ) } . ..., , 2 , , 1 ..., 2 ,
p there are two i, j , , j i ≠
max
1
p Since ,
1
However, we will show that ( )
1 for .
2
C n n n m ..., , , ;
( ) m
max
1
max
1
for ,
2
1
p A
{ } j i n n
i x and . j x
color codes differ in the ordinate of colors
p C But, if they have the same distance, their
≠ because they have different distance to .
Π Π
but they do not have the same distance, then ( ) ( ) v c u c
lie in the same interval, say
max n say max n n i
1
K ,
j n
and
1
K ,
i n
Asmiati 1310
, is ,
= and ,
C n n n m ..., , , ;
K V x ,
( ) m
From all the above cases, we see that the color codes for all vertices in
Π i Π
≠contains exactly one component of value 1. Thus, ( ) ( ) . v c x c
Π
contains at least two components of value 1, whereas ( ) v c
Π
1
∈ and v have the same color, then ( )
i x cj n i
max n n j
K
1 i n
,
not contained in .
1
K ,
i n
< then the color codes of u and v differ in color of leaves of
= | = = |
On the Locating-chromatic Numbers of Non-homogeneous … 1311 Therefore the color codes of and are the same, a contradiction. So,
x x i j
χ ≥ + > (
- C ( m ; n , n , ..., n ) ) n
2 , for p n 1 .
L
1 2 m max max
To show that χ ( C ( m ; n , n , ..., n ) ) ( ≤ n + 2 ) , consider the
L
1 2 m max
locating-coloring c on C ( m ; n , n , ..., n ) as follows:
1 2 m ( ) + • c l = n 2 .
11 max
- The color of vertices x are:
i
1 if i odd , ⎧
=
c ( ) x
⎨
i 2 if i even .
⎩
- Leaves, for n = 1 , c ( ) n =
3 , whereas for n ≥ 2 , { ( ) c a | j = 1 , 2 ,
i i i ij
⊆ + ..., n } , are colored by S { 1 , 2 , ..., n 1 } ( ) \ { c x } for any i.
i max i
Since there is only one vertex at the end of the longest path, which is
- colored by n
2 , the color codes of all vertices are different. Therefore,
max
c is the locating-chromatic coloring on C ( m ; n , n , ..., n ) , and so
1 2 m
χ ( ( + + C m ; n , n , ..., n ) ) ≤ n 2 , for p > n 1 .
L
1 2 m max max
Figure 2. A minimum locating-coloring of C(11; 3, 1, 1, 3, 1, 2, 3, 3, 1, 3, 1).
3. Locating-chromatic Number of Non-homogeneous
Firecracker Graphs
A non-homogeneous firecracker graph, F is obtained by the
n , ( k , k , ..., k )
1 2 n
concatenation of n stars S , i = 1 , 2 , ..., n by linking one leaf from each star.
k i
= | = = − Let
V ( F ) { x , m , l i
1 , 2 , ..., n ; j 1 , 2 , ..., k 2 } ,
n , ( k , k , ..., k ) i i ij i
- n i x x
1 max max max max
number of leaves in the maximal subgraph is ,
− ≤ k p Since the
1 max
L k k k n Proof. We determine the trivial lower bound of .
F n
1 k p if k k p if k
2
..., , , ,
1 , , 1 ,
−
= χ .
− > − ≤ −
⎩ ⎨ ⎧
n is :
( )≥
2
maxF for
2 1 n k k k n
2 max
k by Corollary 1, we have ( ( )
1 and p be the number of subgraphs . max k
1 max
S where . 1 p i
S and denote it by p . Denote each of the subgraphs from left to right by , max i k
a. Find the number of subgraphs max
k
1 as follows:
2
F
..., , , ,k -coloring c on n k k k n
−
p Consider ( )
) ,
− ≤ k
1 max
L k k k n for .
F k n
− ≥ χ
1
2
1 ..., max , , ,
S Then the locating - chromatic number of ( ), ..., , , ,
2
Asmiati 1312 and (
{
max k
= then the subgraph
2 1 max n k k k k
i k j If { } , ..., , , max
1 − =
, 2 ..., 2 ,
l m m x ij i i i ; ..., n } .
, 2 , = 1 , |i
i i ∪
..., , , ,
− = |
1
1
= { } , 1 ..., 2 ,
1
2
F E ..., , , ,
( ) ) n k k k n
S is the maximum star subgraph of non-homogeneous firecracker
( ) .2 1 n k k k n
F ..., , , ,
(1) ( ) ( ) , , , m j m i
S be the maximum star subgraph of ( ) n k k k n
Theorem 3. Let max k
S and j k S are called star subgraphs with the same distance.
S then subgraphs i k
≠ x x x , are the centers of , max k
= x x d x x d with x and p p
S or (2) ( ) ( ) , , , p j i
= x x d x x d with
m
x is the center of , max k≠ = such that
F If there are p subgraphs , max k
, max n n n j i
2 1 n j i k k k n k k F S S ⊂ where . 1 m j i ≤ ≠ ≤ If
..., , , ,
Definition 2. Let ( ), ,
1 p i ≤ ≤
S where .
max i k
left to right, is denoted by ,
S then every subgraph, from
≤ ≤
On the Locating-chromatic Numbers of Non-homogeneous … 1313 i
b. Vertices x ∈ S , where 1 ≤ i ≤ p , are colored by 3 , 4 , 5 , ..., n , i k max
2 , 3 , respectively. i
∈ ≤ ≤
c. Vertices m S , where 1 i p , are colored by 1 , 2 , 3 , ..., p , i k max respectively. i
d. Leaves in S , where 1 ≤ i ≤ p are colored by {
1 , 2 , 3 , ...,
k max
( k 1 )} ( ) \ { c x } .
−
max i i
e. Let A be an open interval before S , A be an open interval
- 1 t
1 k max
- t t
1 between S and S , where 1 ≤ t ≤ p − 1 , and A be an
- p
1 k k max max p open interval after S . k max
f. Let T = {all combinations ( k −
2 ) from ( k − 1 ) color}, such that
max max T = { T , T , ..., T } with T ∈ T be combinations not containing
1
2
1 k − i max color i .
g. Identify subgraph S in the interval as defined in item . d k i
h. If S lies in the interval A or A , then every vertex of S is k
1 2 k i i colored by colors that correspond with T , respectively.
1 3 ≤ k ≤ p
- i. If S lies in the interval A , where
1 , then every vertex k k i of S , is colored by colors that correspond with T , respectively. k k
1 − i
=
j. Let S and S , where k k have the same distance from the k k i j i j
maximum star subgraph. If c ( ) x = c ( ) x , then c ( ) m = c ( ) m .
i j i jNext, we show that the color codes for all vertices in F for n , k , k , ..., k
1 2 n p ≤ k −
1 , are different. Consider two distinct vertices u ∈
V ( S ) and max k i v ∈ V ( S ) , where c ( ) u c ( ) v .
=
k j
1314 Asmiati
= ( ) • If = , then ( ) ≠ because their color codes
k k k c u c v i j max Π Π differ in ordinate of colors m and m . i j
- If S and S lie in different intervals, say A and A , then c ( ) u
k k p q Π i j
≠ c ( ) v because they have different distances from C and C .
Π p q
- If S and S lie in the same interval, say A but they do not have
k p k i j the same distance, then c ( ) u ≠ c ( ) v because they have different
Π Π
distance to C . But, if they have the same distance, their color codes p differ in ordinates m and m . i j- If one of k , k } is k , say k = k and n k , then the
<
{
i j max i max j max color codes of u and v differ in color of leaf of S , not contained in k i S . k j
- ∈
If x V ( K ) and v lie in the same interval, say A , where c ( ) x i 1 , n k i j
= c ( ) v , then c ( ) x and c ( ) v have different distance from C . So,
Π i Π k c ( ) x c ( ) v . But, if they lie in different intervals, say A and A ,
≠
i r s
Π Π then c ( ) x c ( ) v because they have different distances from C
≠
Π i Π r and C . s
From all above cases, we see that the color codes for all vertices in F , for p ≤ k − 1 are different, thus c is a locating-coloring. n , k , k , ..., k max
1 2 n So, χ ( F ) ≤ k − 1 , for p ≤ k − 1 .
L n , ( k , k , ..., k ) max max
1 2 n
> k −
We will show the lower bound for p 1 . By Corollary 1, we max
χ ≥ −
have that ( F ) k 1 . However, we will show that k L n , ( k , k , ..., k ) max max
1 2 n
−
1 colors are not enough. For a contradiction that there exists a ( k − 1 ) - max locating coloring c on F for p > k − 1 . Since p > k n , ( k , k , ..., k ), max max
1 2 n
[3] Asmiati and E. T. Baskoro, Characterizing all graphs containing cycle with the
locating-chromatic number 3, AIP Conf. Proc. 1450, 2012, pp. 351-357.[4] Asmiati, H. Assiyatun and E. T. Baskoro, Locating-chromatic number of
amalgamation of stars, ITB J. Sci. 43A (2011), 1-8.[5] E. T. Baskoro and Asmiati, Characterizing all trees with locating-chromatic
max A k
. otherwise , 2 \ , , 1 if
= = − = |
⎩ ⎨ ⎧
{ }
define: { ( ) } { }
=
1
1 max max
i
m c for i otherwise.
2 =
1 = k m c and ( )
max
i x c if i is even.
=
1 \ , 2 ..., 2 ,
A k A i k k j l c i ij It is easy to verify that the color codes of all vertices are different.
i x c if i is odd and ( )
1 F k n
Uttunggadewa, The locating-chromatic number of firecracker graphs, Far East J. Math. Sci. (FJMS) 63(1) (2012), 11-23.
[2] Asmiati, E. T. Baskoro, H. Assiyatun, D. Suprijanto, R. Simanjuntak and S.
References
[1] Asmiati, The locating-chromatic number of non-homogeneous amalgamation of
stars, Far East J. Math. Sci. (FJMS) 93(1) (2014), 89-96.max − > k p This completes the proof.
1
L k k k n ≤ χ for .
2
Therefore, c is a locating-chromatic coloring on ( )
..., max , , ,
( ) ) ,
F and so (
2 1 n k k k n
..., , , ,
,
3
=
On the Locating-chromatic Numbers of Non-homogeneous … 1315 , 1 − there are two i
k l l c jl
≥ χ for .
L k k k n
1
F k n2
) ..., max , , ,
Therefore, the color codes of i m and j m are the same, a contradiction. So, ( ( )
− = |
− > k
1 max
2 ,
1 max k h l c ih { ( ) } . , 2 ...,
≠ such that ( ) { } = − = |
, 2 ..., 2 ,, , j i
, j
1 max
p Next, we determine the upper bound of ( ) n k k k n
1
1 F k n
1 as follows:
2
F ..., , , ,
on ( ) n k k k n
≤ χ consider the locating-coloring c
L k k k n
2
F ..., , , ,
..., max , , ,
To show that ( ( )
) ,
− > k p
max
1
1 for .
2
- ( )
- ( )
- If { } , ..., , 2 ,
Asmiati 1316
[6] E. T. Baskoro and I. A. Purwasih, The locating-chromatic number for corona
product of graphs, AIP Conf. Proc. 1450, 2012, pp. 342-345.
[7] A. Behtoei and B. Omoomi, On the locating chromatic number of Kneser graphs,
Discrete Appl. Math. 159 (2011), 2214-2221.[8] G. Chartrand, D. Erwin, M. A. Henning, P. J. Slater and P. Zhang, Graphs of
order n with locating-chromatic number n – 1, Discrete Math. 269(1-3) (2003),65-79.