Independence Related Parameters and Edge Removal
D. K. THAKKAR
1and D. D. PANDYA
21
Department of Mathematics,
Saurashtra University, Rajkot, Gujarat, INDIA.
2
Mathematics Department, L. E. College, Morbi, Gujarat, INDIA.
(Received on: May 9, 2013) ABSTRACT
Independence, Independent Domination and Vertex Covering are Parameters related to Domination. In this paper we establish necessary and sufficient conditions under which these parameters change when an edge is removed from the graph. In particular We prove that the vertex covering number of the graph does not increase when an edge is removed from the graph. we give some related examples.
Keywords: Independent Dominating Set, Independent Domination Number, Minimal Independent Dominating Set, Independence Set, Maximum Independence Set, Vertex Covering Set, Minimum Vertex Covering set.
AMS subject classification (2000): 05c69.
PRELIMINARIES AND NOTATIONS
If G is a graph then V(G) or V will denote the set of vertices of the graph G And E(G) or E will denote the set of edges of the graph G. If u and v are the end vertices of an edge e then we will write e=uv or e=vu. The sub graph obtained by removing an edge e=uv will be denoted as G - {e} or G - {uv}.The sub graph obtained by removing a vertex v from the graph G will be denoted as G - {v}.we will consider only simple and undirected graphs in this paper.
Vertex covering
Definition 1.1 Vertex Covering Set[6]
Let G be a graph. A set S ⊂ V G ( ) is said to be Vertex Covering Set of the graph G if every edge has at least one end point in S.
Definition 1.2 Minimum Vertex Covering Set[6]
A Vertex Covering Set with minimum cardinality is called minimum vertex covering set. It is also called γ
crSet.
Definition 1.3 Vertex Covering
Number[6]
The Vertex Covering Number of the graph G is the Cardinality of any minimum Vertex Covering Set of the graph G. It is denoted by α
0( ) G or simply α
0.
>>It is useful to partition the Edge set of G into two sets according to how their Removal affects γ
cr( ) G .
0
{ ( ) / ( ) ( )}
{ ( ) / ( ) ( )}
cr cr cr
cr cr cr
E e E G G e G
E e E G G e G
γ γ
γ γ
−
= ∈ − <
= ∈ − =
It is Obvious that E = E
cr−∪ E
cr0>>It is useful to partition the Vertex set of G into two sets according to how their Removal affects γ
cr( ) G .
0
{ ( ) / ( ) ( )}
{ ( ) / ( ) ( )}
cr cr cr
cr cr cr
V v V G G v G
V v V G G v G
γ γ
γ γ
−
= ∈ − <
= ∈ − =
It is Obvious that V = V
cr−∪ V
cr0The following theorem says that the vertex covering number of the graph does not increase when an edge is removed from graph.
Theorem 1.4 Let G be a graph and e be an edge of G. Then α
0( G − < e ) α
0( ) G
Proof: Let S be a minimum vertex covering set of G. It is obvious that S is also a vertex covering set of G − { } e Therefore
0
( G e ) S
0( ) G α − < = α .
This proves the theorem. █
Theorem 1.5: An Edge
e = uv ∈ E
cr−⇔ There is a minimum vertex covering Set S of G
such that All neighbours of u except possibly v are in S and
All neighbours of v except possibly u are in S.
Proof: ⇒
Suppose e = uv ∈ E
cr−Let S
1be a minimum vertex covering set of { }
G − uv .Now S
1can not be a Vertex Covering Set of G. Therefore u S ∉
1& v S ∉
1. Let S = ∪ S
1{ } u . Then because of condition given , S is a minimum vertex Covering Set of G.
Now if we consider S then u S ∈ . Let w be any neighbour of u in G − { } uv . Since S
1is a Vertex Covering Set of G − { } uv , w S ∈
1and hence w S ∈ . So, All neighbours of u except v are in S & All neighbours of v are in S.
⇐ Conversely
Let S be a minimum Vertex Covering Set of G such that u S ∈ . Let
1
{ }
S = − S u . Since every neighbour of u except possible v are in S. It follows that every neighbour of u in G − { } uv is in S.
If v ∉ S then every neighbour of v in { }
G − uv is in S
1.Thus S
1is Vertex
Covering Set in G − { } uv . █
Theorem 1.6: An Edge e = ∈ uv E
cr0⇔ For every minimum Vertex Covering Set S of G
There is a neighbour of u different from v which lies outside of S OR
There is a neighbour of v different from u which lies outside of S.
Proof: ⇐
Suppose Condition holds.
Let T be a minimum Vertex Covering Set of { }
G − uv .
Claim: | T | = γ
cr( ) G
Suppose | T | < γ
cr( ) G . Then T can not be Vertex Covering Set of G. Therefore
&
u T ∉ v T ∉ .
Now Consider S = ∪ T { } u .Obviously S is a minimum Vertex Covering Set of G. Since T is Vertex Covering Set of G − { } uv . Every neighbour of u in G − { } uv is in T.
Therefore every neighbour of u different from v in G lies in S.
This Contradicts the hypothesis.
Therefore | T | = γ
cr( ) G and e = ∈ uv E
cr0.
⇒ Suppose
0e = uv ∈ E
crLet S be a minimum Vertex Covering Set of G. Then u S or v ∈ ∈ S
Suppose Every neighbour of u different from v is in S and Every neighbour of v different from u is in S. Then S − { } u is a
Vertex Covering Set of G − { } uv which
implies that γ
cr( G uv − ) < γ
cr( ) G Which is a Contradiction.
So condition follows. █
Lemma 1.7: Let G be a Graph and e=uv be an edge of Graph G. Then
0
( G v )
0( G uv )
0( ) G α − < α − < α .
Proof: We need to prove only that
0
( G v )
0( G uv ) α − < α −
Let S be a minimum vertex covering set of { }
G − uv
Case 1: u S ∉ and v S ∉
Obviously S is a vertex covering set of { }
G − v
Case 2: u S ∈ and v S ∉
Here also S is a vertex covering set of { }
G − v .
Case 3: u S ∉ and v S ∈
If xy is an edge of G − { } v then and
x ≠ v y ≠ v .
If x=u then uy is an edge of G − { } uv and therefore it has an end vertex in S.
Similarly y=u then ux is an edge of { }
G − uv and therefore it has an end vertex in S.
Case 4: u S ∈ and v S ∈
Consider the set S
1= − S { } v . If xy is an
edge of G − { } v then x ≠ v and y ≠ v . Then
xy is an edge of G − { } uv not containing the
vertex v. Since S is a vertex covering set of { }
G − uv , x S or y S ∈ ∈ Equivalently
1 1
x S or y S ∈ ∈ . Thus S
1is a vertex covering set of G − { } v .
Hence α
0( G v − < ) α
0( G uv − ) . █ Theorem 1.6 : Let G be a Graph and e=uv be an edge of Graph G. If uv E ∈
cr−then
cr
and
cru V ∈
−v V ∈
−.
Proof: Suppose uv E ∈
cr−0 0
Then α ( G uv − ) < α ( ) G .
And we knowthat α
0( G v − < ) α
0( G uv − ) .
Therefore α
0( G v − < ) α
0( ) G and Similarly α
0( G u − < ) α
0( ) G . Hence
cr
and
cru V ∈
−v V ∈
−. █
*Following Example shows that converse of above theorem is not always true.
Example 1.7: Consider the graph P
6(path with 6 vertices).Here u V ∈
cr−and v V ∈
cr−but e = ∉ uv E
cr−Infact e = ∈ uv E
cr0.Example 1.8: Consider the Graph P
5with vertex set S={a,b,u,v,c}.It can be verified that the edge uv E ∈
cr0.
Example 1.9: Consider the Peterson graph. In Peterson graph all edges are in
0
E
cr. Here |S|=6. Consider the edge e=uv.
We can see that all neighbours of u are in S.
Note: It may be noted that every vertex in Peterson graph is in V
cr−.
u v
a b u v c a b u v c
u
v u
v
Example 1.10: Consider the Graph C
3with vertex set S={a,u,v}. It can be verified that the edge uv E ∈
cr−.
*In [2] it has been proved that (1) A vertex v V ∈
cr−⇔ there is a minimum vertex covering set S such that v ∈ S . (2) A vertex
0
v V ∈
cr⇔ v ∉ S for every minimum vertex covering set S of G.
It has been further proved that if
0
v V ∈
crthen for every vertex u such that u is adjacent to v, u V ∈
cr−Infact every neighbour of v belongs to every minimum vertex covering set of G.
That is N v ( ) is subset of intersection of all minimum vertex covering sets of G.
Corollary 1.11: If v V ∈
cr0then for every edge e = uv , e = ∈ uv E
cr0Proof : Since v V ∈
cr0. α
0( G v − = ) α
0( ) G . Now from Lemma 1.7
0
( G v )
0( G uv )
0( ) G
α − = α − = α
Hence e = ∈ uv E
cr0█ MAXIMUM INDEPENDENT SET
Definition 2.1 Independent Set[6]
Let G be a graph. A set S ⊂ V G ( ) is said to be Independent Set if any two distinct vertices of S are non-adjacent.
Definition 2.1 Maximum Independent Set[6]
An Independent Set with maximum cardinality is called maximum Independent Set.
Definition 2.3 Independence Number[6]
The cardinality of a maximum Independent set is called Independence Number of the graph G and it is denoted by
0
( ) G
β or simply β
0.
>>It is useful to partition the Edge set of G into two sets according to how their Removal affects β
0( ) G .
0 0
0
0 0
{ ( ) / ( ) ( )}
{ ( ) / ( ) ( )}
I I
E e E G G e G
E e E G G e G
β β
β β
+
= ∈ − >
= ∈ − =
It is Obvious that E = E
I+∪ E
I0>>It is useful to partition the Vertex set of G into two sets according to how their Removal affects β
0( ) G .
0 0
0
0 0
{ ( ) / ( ) ( )}
{ ( ) / ( ) ( )}
I I
V v V G G v G
V v V G G v G
β β
β β
−
= ∈ − <
= ∈ − =
It is Obvious that V = V
I−∪ V
I0a a
u v u v
It may be noted that the independence number of the graph does not decrease when an edge is removed from the graph.
In the following theorem we prove the necessary and sufficient condition under which this number increases when an edge is removed.
Theorem 2.4: Let G be a graph and e=uv be an edge of graph G. Then
0
( G e )
0( ) G β − > β ⇔
There is a maximum independence set S of G such that
, & ( ) { } u ∈ S v ∉ S N v ∩ = S u Proof:
⇐ (sufficiency)
Let S be a maximum independence set of G such that u ∈ S v , ∉ S & N v ( ) ∩ = S { } u .Now v is not adjacent to any vertex of S in graph G-{e} and therefore S ∪ { } v is an independence set in G-{e}.
Thus β
0( G e − > ∪ ) | S { } | v > β
0( ) G .
⇒ (Necessity)
Suppose β
0( G e − > ) β
0( ) G
Let S be any maximum independence set in G-{e}.Then S can not be independence set in G. This implies that u v , ∈ S .
Now Consider the set S-{v} in graph G.
Since S is independence in G-{e}, v is adjacent to only one vertex u of S- {v}.Therefore N v ( ) ∩ − S { } { } v = u .Also
{ }
S − v is independence set in G-{e}
because v
∉SIf T is an independence set in G such that
| T | | > − S { }| v then | | | | T > S .This implies that β
0( ) | G > T | > β
0( G e − ) Which is a contradiction.
So, S-{v} is a maximum independence set in G. Thus condition is satisfied by S-{v}. This completes the proof. █
Theorem 2.5: Let G be a graph and e=uv be an edge of graph G then
0
( G e )
0( ) G
β − = β ⇔ (for every
maximum independence set S of G) one of following holds
(1) u ∈ S v , ∉ S &| N v ( ) ∩ > S | 2
(2) u v , ∉ S
Proof:
⇐ CASE A: Let S be maximum independence set of G and
, &| ( ) | 2 u ∈ S v ∉ S N v ∩ > S .
Consider G-{e}.Now S is still independence set of G-{e}. Let T be maximum independence set of G-{e} with
|T|>|S|
Case (1) u ∈ T v , ∉ T T is maximum independence set of G with |T|>|S|.
So, β
0( ) | G > T | | > = S | β
0( ) G . Which is a contradiction.
Case (2) u v , ∉ T .T is maximum independence set of G with |T|>|S|.
So, β
0( ) | G > T | | > = S | β
0( ) G .Which is a contradiction
Case (3) u v , ∈ T Consider T-{v}. So T-
{v} is maximum independence set of G with
|T-{v}|(maximal)=|S|(maximum) Thus T-{v} is maximum independence set in G such that v T ∉ − { } v and N v ( ) ∩ − T { } v contains exactly one vertex u , which is a contradiction. Therefore such a set T does not exist in G-{e}.
Therefore S is a maximum independence set n G-{e}.
CASE B: Suppose u v , ∉ S
If then according to previous theorem, there must be a maximum independence set S
1such that
1
,
1& ( )
1{ }
u ∈ S v ∉ S N v ∩ = S u .Which is not true in this Case.
Therefore β
0( G e − = ) β
0( ) G
⇒ (Necessity)
Let S be a maximum independence Set in G such that β
0( G e − = ) β
0( ) G
Suppose (1) u v , ∉ S then condition (2) is satisfied.
Suppose(2) u v , ∈ S . which is not possible as S is maximum independence set.
Suppose(3) u ∈ S v , ∉ S & N v ( ) ∩ = S { } u Here v is not adjacent to any vertex of S in G-{e}.Therefore S ∪ { } v is an independence set in G-{e}.
Thus | S ∪ { }| | v > S |
So, β
0( G e − > ∪ ) | S { }| | | v > = S β
0( ) G . Which is a contradiction.
So, u ∈ S v , ∉ S &| N v ( ) ∩ > S | 2 █
*In [3] it has been Proved that a vertex
0
v V ∈
I⇔ There is a maximum independent set S of G such that v ∉ S
Corollary 2.6: Suppose G is a graph and e = uv be an edge of graph G.
e = ∈ uv E
I+then u V ∈
I0or v V ∈
I0Proof: By above theorem ,There is a maximum independent set S such that
, and ( ) { }
u S v S ∈ ∉ N v ∩ = S u
By above remark, it is clear that v V ∈
I0. Corollary 2.7: Let G be a Graph and
e = uv be an edge. If e = ∈ uv E
I0then
0 0
I
or
Iu V ∈ v V ∈
In [3] it has been proved that Suppose v V ∈
I0then
(1) If | N v ( ) ∩ > S | 2 then all edge belongs to E
I0(2) If | N v ( ) ∩ = S | { } u then e = ∈ uv E
I+.
Corollary 2.8:
Suppose v V ∈
I0then the edge e = ∈ uv E
I+⇔ there is a maximum independence set S of G such that u S ∈ and N v ( ) ∩ = S { } u Note: If v V ∈
I0then all edges incident to v are in E
I0Corollary: Let G be a Graph and e=uv . If β
0( G v − = ) β
0( ) G then for every edge
0 0
, ( ) ( )
uv β G uv − = β G
INDEPENDENT DOMINATION
Definition 3.1 Independent Dominating
Set
Let G be a graph. A set S ⊂ V G ( ) is said to be Independent Dominating Set of the graph G if (1) S is Dominating Set and (2) S is an Independent Set.
Definition 3.2 Minimum Independent Dominating Set
An Independent Dominating Set with minimum cardinality is called minimum Independent Dominating set.
Definition 3.3 Independent Domination Number
The Independent Domination Number of the graph G is the Cardinality of any minimum Independent Dominating Set of the graph G denoted by i G ( )
>>It is useful to partition the Edge set of G into two sets according to how their Removal affects i G ( ) .
0
{ ( ) / ( ) ( )}
{ ( ) / ( ) ( )}
{ ( ) / ( ) ( )}
i i i
E e E G i G e i G E e E G i G e i G E e E G i G e i G
−
+
= ∈ − <
= ∈ − =
= ∈ − >
It is Obvious that E = E
i−∪ E
i0∪ E
i+Theorem 3.4: An edge e = P P
1 2∈ E
i +⇔ (for every γ
iset S of G) P
1∈ S P ,
2∉ S and for every w ≠ P
1in N P ( )
2, N w ( ) ∩ ≠ S φ
Proof:
⇐ Let S be a γ
iset in G and condition satisfies.
Consider G – {e}. Now S is not independent dominating set in G – {e}.
Let T be γ
iset in G – {e} with | | | | T < S
(a) P
1∈ T P ,
2∈ T . Now in G, T − { } P
2will be γ
iset with | T − { } | | P
2< S | .
which is a contradiction.
(b) P
1∉ T P ,
2∉ T .Now in G, T is γ
iset with | | | | T < S and P or P T
1 2∉ which contradicts our assumption.
(c) P
1∈ T P ,
2∉ T .Now P
2is adjacent to x T ∈ .Consider G,Now T is independent dominating set in G also. But N x ( ) ∩ = S φ
because x T ∈ . Which contradicts our assumption.
So, e = P P
1 2∈ E
i +⇒ An edge
1 2 i
e = P P ∈ E
+Let S be any γ
iset of G. Then S can not be a dominating set in G – {e}.There are two possibilities.
(1) P
1& P
2∉ S .Then S is a dominating set in G – {e}. Which is a contradiction.
(2) P S P
1∈ ,
2∉ S .Now S is not a dominating set in G – {e}.Therefore P
2is not adjacent to any vertex of S. If
2 1
( ) &
w N P ∈ w ≠ P then w is adjacent to some vertex of S.
Therefore N w ( ) ∩ ≠ S φ █
Theorem 3.5: An Edge e = P P
1 2∈ E
i −⇔ (For any dominating set T of G with
| T | < γ
i( ) G ) P
1& P
2∈ T and they are the only non-isolated vertices in T.
Proof:
⇐ Let T be a set which satisfies given conditions in Theorem. Then T is a independent dominating set in G − { P P
1 2} Therefore i G ( − P P
1 2) | < T | < i G ( ) . Therefore e = P P
1 2∈ E
i −.
⇒ Suppose
1 2 i
e = P P ∈ E
−Let T be a minimum independent dominating set in G − { P P
1 2} .Then T can not be independent dominating set in G.
Since T must be a dominating set in G, it can not be independent set in G.
Therefore P
1& P
2∈ T and Obviously they are the only non-isolated vertices in T. █
Note: In Peterson graph, wheel graph all
0
E ∈ E
i.
REFERENCES
1. C.L. Suffel, D. Bauer, F. Harary, J.
Nieminen. : Domination alteration sets in graphs. : Discrete Math. 47:153-161, (1983).
2. D.K. Thakkar and J.C. Bosamiya.:
Graph Critical With Respect To Variants of Domination, (2011).
3. D. K. Thakkar and J. V. Changela,:
Mathematical Modelling, (2011).
4. J.F. Fink, J. Roberts, M.S. Jacobson, L.F. Kinch. On graphs having domination number half their order.
Period. Math Hungar. 16:287-293, (1985).
5. H.B. Walikar and B.D. Acharya.:
Domination in critical Graphs: Nat.
Acad. Sci. Lett.2:70-72, (1979).