1. Introduction
In graph protection, mobile agents or guards are placed on vertices in order to defend against a sequence of attacks on a network. See [1] [2] [3] [4] [5] for more background of the graph protection problem. The first idea for eternal domination was introduced by Burger et al. in 2004 [1]. The “all guards move model” or “multiple guards move version” of eternal domination was introduced by Goddard et al. [2]. General bounds of
were determined in [2], where
denotes the domination number of G and
denotes independence number of G. The eternal domination number for cycles
and paths
was found by Goddard et al. [2] as follows:
and
. Jahangir Graph
for
is a graph on
vertices,
i.e. a graph consisting of a cycle
with one additional vertex which is adjacent to m vertices of
at distance s from each other on
see [6] for more information on Jahangir graph. Let
be the label of the central vertex and
be the labels of the vertices that incident clockwise on cycle
so that
. We will use this labeling for the rest of the paper. The vertices that are adjacent to
have the labels
. We denote the set
by R. So,
. By definition, for
, Jahangir Graph
is the wheel graph
and it was mentioned in [6] that
for
. The k-dominating graph
was defined by Goldwasser et al. [7] as follows: Let G be a graph with a dominating set of cardinality k. The vertex set of the k-dominating graph
, denoted
, is the set of all subsets of
of size k which are dominating sets and two vertices of H are adjacent if and only if the k guards occupying the vertices of G of one can move (at most distance one each) to the vertices of the other,
if and only if
has an induced subgraph
such that for each vertex x of
, the union of the vertices in the closed neighborhood of x in
is equal to
.
Proposition 1.1 [6] :
for
.
2. Main Results
Eternal Domination Number of
In this section, we give the exact eternal domination number of
.
Lemma 2.1: Let us have a graph
. For
when m is even and
when m is odd, then a set
of cardinality
can’t dominate
if
.
Proof: Since
that means all the vertices of S are vertices from the
outer cycle
. We know that
. So let’s find out the values of m for which:
. This arbitrator is true for m is even with
and for m is odd with
. ■
Theorem 2.1.
Proof: We know from the definition of eternal domination that
.
Therefore from proposition 1.1, we have
for
. This
means we only need to prove that
for
. In order to do that we form the k-dominating graph
on graph
with
and
. We consider the following cases:
Case 1.
: It was found in [3] that
. Therefore
. However, it is obvious that two vertices can dominate
if and only if both vertices belong to the outer cycle
. Therefore if the central vertex
is attacked, then one of the two guards that are located on the two dominating vertices would have to move to
making it impossible for the new distribution of guards to dominate the entire graph because
doesn’t belong to any of the 2-dominating sets of
. Therefore
. We form
, the 3-dominating graph on
. Let
be the induced subgraph of
on vertices:
,
,
. Since
are all adjacent and
, therefore we have
, which means
, therefore
.
Case 2.
: We have
. Let’s form
to be the induced subgraph of
on vertices
,
,
. Since
are adjacent and
therefore we have
which means
.
Case 3.
: We have
. Let’s form
to be the
induced subgraph of
on the following vertices:
,
,
,
. Since
are all adjacent and
, therefore
which means
.
Case 4.
: We have
. Let’s form
to be the
induced subgraph of
on the following vertices:
,
,
,
. Since
are all adjacent and
therefore
which means
.
Case 5.
: We have
. Let’s form
to be the
induced subgraph of
on these vertices:
,
,
,
. Since
are adjacent and
, therefore
which means
.
Case 6.
: We have
. Let’s form
to be the induced subgraph of
on the following vertices:
,
,
,
. Since
are adjacent and
, therefore
which means
.
Case 7.
: We have
. Let’s form
to be the
induced subgraph of
on the vertices:
,
,
,
. Since
are all adjacent and
, therefore
which means
.
(See Figure 1 for
). ■
Lemma 2.2:
for
and
.
Proof: From the definition of eternal domination, we already know that
. By proposition 1.1,
for
. We just need to prove that
for
and
. We consider both cases:
Case 1: m is even for
.
In this case the sets
and
are the only two minimum dominating sets (γ-dominating set) of
where both
and
are similar by symmetry. We study an arbitrary attack on a vertex
from a graph
protected by
and we prove that
fails to eternally protect
. Let the attacked vertex
have an odd (index) label,
. The only guard protecting
in this case is the guard occupying the central vertex
(which is adjacent to all the odd vertices of
). This means the guard on
has to move to
to defend the attack. However, that would leave the vertices:
unprotected. To try to avoid that we have two strategies:
Strategy 1: We move another guard (occupying an odd vertex
:
) to
to keep
protected. However, that would leave at least one of the two vertices
unprotected and this strategy fails, see Figure 2.
Strategy 2: We don’t move any other guard to
which would leave
guards on the vertices of cycle
to protect
. By Lemma 2.1
Figure 2. Illustrating strategy 1 when m = 8.
these guards can’t protect
if
therefore this strategy fails as well, see Figure 3.
Since both of these strategies fail then
for
&
. Without loss of generality, the same argument can be followed to
prove that
guards can’t eternally protect
in case the minimum dominating set is
.
Case 2: m is odd for
.
In this case the minimum dominating sets (γ-dominating sets) of
are:
Figure 3. Illustrating strategy 2 when m = 8.
We study an arbitrary attack on a vertex
from three cases of
protected by
of
respectively. We prove that
fail to eternally protect these graphs. Let the attacked vertex
have an odd (index) label,
. The only guard protecting
in this case is the guard occupying the central vertex
(which is adjacent to all the odd vertices of
). This means the guard on
has to move to
to defend the attack. However, that would leave the vertices:
unprotected. To try to avoid that we have two strategies:
Strategy 1: We move another guard (occupying an odd vertex
) to vertex
to keep
protected. However, that would leave at least one of the two neighboring vertices to
unprotected therefore this strategy fails, see Figure 4.
Strategy 2: We don’t move any other guard to
which would leave
guards on the vertices of cycle
to protect
. By Lemma 2.1 these guards can’t protect
if
, therefore this strategy fails as well, see Figure 5.
Since both strategies fail then
for m is odd and
.
Without loss of generality, the same argument can be followed to prove that
guards cannot eternally protect
in case the minimum dominating set is
or
. From cases 1 and 2 we conclude that:
for
and
.
However, we know
for
, therefore:
for
and
. ■
Theorem 2.2:
for
and
.
Proof: From Lemma 2.2 It is enough to prove the existence of one eternal
dominating family of the vertices of
with cardinality
in order to prove that
. We consider both cases:
Case a. m is even:
We start by forming the k-dominating graph denoted
on
with
.
is a dominating set of
. We form
Y the family of dominating sets as follows
:
. Hence the cardinality of
is
. Therefore each set of the family Y is a vertex of
. It is obvious that the union of these vertices is
. We now need to prove that these vertices are all adjacent in
. There are two types of sets
depending on the label of the vertex
:
Type 1:
.
Type 2:
.
When an arbitrary unoccupied vertex
is attacked we consider the following cases:
Case a.1.
: we consider the following cases:
Case a.1.1.
is an unoccupied odd vertex
: In this case the guard on the central vertex
moves to
to defend the attack and the guard on
moves to
to protect the remaining odd vertices without disturbing the protection of the even vertices (which are protected by the
guards on the vertices of
) (see Figure 6), which means
guards are
enough to protect
in this case. After defending the attack and since
the resulting dominating set
. We will now use the same argument to prove the results for all cases of
when m is even, taking into consideration that the same path can be followed to prove all the possible cases of
when m is odd.
Case a.1.2.
is an even vertex
: In this case the neighboring odd vertex has the only available guard to defend
. So the guard on
(or
) moves to
to defend the attack leaving
(or
) respectively unprotected, so the guard on
moves to
(or
) respectively, and the guard on
moves to
to protect the remaining vertices of M. While the guards on the vertices of the set
keep protecting those vertices and the even vertices of
leaving
protected, see Figure 7.
After defending the attack and since
the resulting dominating set
.
Case a.2:
, we consider the following cases:
Case a.2.1:
is an unoccupied odd vertex
. In this case the guard on
moves to
, either
or
. So the guard on
(or
) moves to
and the guard on
moves to
(or
) respectively, keeping the entire graph protected. After defending the attack and since
the resulting dominating set
. Figure 8, illustrates the
process in which
guards can successfully defend
when the
attacked vertex
has an odd index label (
) while the additional guard besides
has an even index label (
).
Case a.2.2:
is an even index vertex
: In this case
, so the guard on
(or
) moves to
. The guard on
moves to
(or
) respectively. The guard on
(or
) moves to
and the guard on
moves to
(or
) respectively leaving the graph
fully protected, see Figure 9.
After defending the attack and since
the resulting dominating set
.
After discussing all possible cases we find that for any
:
are
adjacent in
because the guards occupying
can move to occupy
in one move and vice versa. Therefore we form
on
the vertices
, therefore these vertices are adjacent in the induced subgraph
. It is obvious that
, therefore
for m is even and
. However, from Lemma 2.2
. Therefore
for m is even and
.
Case b. m is odd and
:
We begin by forming the k-dominating graph
on
with
.
is a dominating set of
. We
form Y the family of dominating sets
:
.
Hence the cardinality of
is
. Therefore each set of the family Y is a vertex of
. It is obvious that the union of these vertices is
. We now need to prove that these vertices are all adjacent in
.
There are two types of
dependeng on the label of the vertex
:
Type 1: O = {
where
and
is an unoccupied odd vertex of
}.
Type 2: Q = {
where
and
is an even vertex of
}.
By following the same argument that we followed in case a, we conclude that:
for m is odd and
.
From case a and case b we conclude that:
for
and
. ■
3. Domination and Eternal Domination Numbers of
In this section we consider the graph
. So, we found the exact domination and eternal domination numbers of
.
Theorem 3.1:
for
and the γ-dominating set is unique.
Proof: For
, let
. Since
it is easy to verify that the set of vertices
is a dominating set of cardinality m for
. Therefore
for
. Let
such that
and
is a dominating set of
. We consider the following cases:
Case a: Let
then the vertex
clearly dominates m vertices of the cycle
, which leaves the remaining
guards in the set
to dominate the remaining 2m vertices of cycle
.
are m subsets of cardinality 2, each consists of two non-dominated vertices of
. In order to dominate each of these subsets we need a vertex
, which means we need at least m vertices to dominate these remaining vertices. Therefore since
, there are at least four vertices of
that no vertices of
can dominate which is a contradiction.
Case b:
. In this case
is a dominating set of
vertices that
dominates a cycle
which creates a contradiction since
.
Therefore,
. Finally, by case a and case b we conclude that
is the unique dominating set of cardinality m for
. ■
Lemma 3.2:
for
.
Proof: In Theorem 3.1, we found that
. Since
, we conclude that
. Let
be the m-dominating set of
and let’s assume that each vertex of
is occupied by a guard. When an unoccupied vertex
is attacked we consider the following cases:
Case a. The attacked vertex
: In this case the only guard that can move to
to defend the attack is
or
because
.
Case a.1: If the guard is situated on
then it moves to
to defend the attack. Therefore all the guards of the exterior cycle
should move one edge (counter clockwise) to keep the cycle protected making the new dominating set
. However, according to Theorem 3.1 the vertex
won’t be protected anymore, see Figure 10.
Case a.2: If the guard is situated on
then it moves to
and all the guards on the vertices of
should move one edge clockwise to keep the cycle
protected making the new dominating set
. However, according to Theorem 3.1 the vertex
won’t be protected anymore, see Figure 11.
Case b: The attacked vertex
is
. In this case, there are m guards on the vertices of
each one qualifies to move to
. Let
have the guard that moves to
. This leaves the two vertices
unprotected and there are no available guards on the cycle
to protect them without leaving gaps of unprotected vertices. From cases a and b we conclude that
. Hence
. ■
Theorem 3.3:
for
.
Proof: We form the
(k-dominating Graph) on
with
. Let the dominating sets
be defined as follows:
Each of
is a dominating set of the cardinality
for
therefore they are vertices of
and they are all adjacent in
because they are reachable from each other in one step only. With the guard on the central vertex
staying in place, the dominating sets
can result from each other as follows:
We form
the induced subgraph from
on the previous vertices
. Since
then
.
Now, by last results together with Lemma 3.2 that
. Therefore
. See Figure 12. ■
4. Conclusion
In this paper, we studied the eternal domination number of Jahangir graph
for =2, 3 and arbitrary m. We also find the domination number for
. By using the same approach, we will work to find the eternal domination number of Jahangir graph
for arbitraries s and m.