NOTE ON THE SPLIT DOMINATION NUMBER OF THE CARTESIAN PRODUCT OF PATHS
Maciej Zwierzchowski Institute of Mathematics Technical University of Szczecin Al. Piast´ow 48/49, 70–310 Szczecin, Poland
e-mail: mzwierz@ps.pl
Abstract
In this note the split domination number of the Cartesian prod- uct of two paths is considered. Our results are related to [2] where the domination number of P
m¤P
nwas studied. The split domination number of P
2¤P
nis calculated, and we give good estimates for the split domination number of P
m¤P
nexpressed in terms of its domina- tion number.
Keywords: domination number, split domination number, Cartesian product of graphs.
2000 Mathematics Subject Classification: 05C69.
1. Introduction
In this paper we consider finite undirected simple graphs. For any graph G we denote V (G) and E(G), the vertex set of G and the edge set of G, respectively. If n is the cardinality of V (G), then we say that G is of order n.
By hXi
Gwe mean a subgraph of a graph G induced by a subset X ⊆ V (G).
A subset D ⊆ V (G) is a dominating set of G, if for every x ∈ V (G)−D, there
is a vertex y ∈ D such that xy ∈ E(G). We also say that x is dominated by
D in G or by y in G. A dominating set D of G is a split dominating set of G,
if the induced subgraph hV (G) − Di
Gof G is disconnected. The domination
number, [the split domination number] of a graph G, denoted γ(G), [γ
s(G)] is
the cardinality of the smallest dominating [the smallest split dominating] set
of G. A dominating set D is called a γ(G)-set if D realizes the domination number. Similarly we define a γ
s(G)-set. From the definition of a split dominating set it follows immediately that γ(G) ≤ γ
s(G). Additionally note that for a connected graph G a γ
s(G)-set exists if and only if G is different from a complete graph. More information about a split dominating set and the split domination number can be found in [3]. The Cartesian product of two graphs G and H, is a graph G¤H with V (G¤H) = V (G) × V (H) and (g
1, h
1)(g
2, h
2) ∈ E(G¤H) if and only if (g
1= g
2and h
1h
2∈ E(H)) or (g
1g
2∈ E(G) and h
1= h
2).
Any other terms not defined in this paper can be found in [1].
2. Main Results
Theorem 1. For any n, m ≥ 2
γ(P
m¤P
n) ≤ γ
s(P
m¤P
n) ≤ γ(P
m¤P
n) + 1.
P roof. Let m, n ≥ 2 and let D be the minimum dominating set of P
m¤P
n. According to the definition of a split dominating set we have γ(P
m¤P
n) ≤ γ
s(P
m¤P
n). Thus to prove this theorem we will show that γ
s(P
m¤P
n) ≤ γ(P
m¤P
n) + 1. Consider the graph P
m¤P
n, as m canonical copies of P
nwith vertices labelled x
i,j, for i = 1, 2, . . . , n and j = 1, 2, . . . , m, and with edges x
i,jx
i+1,jand x
i,jx
i,j+1.
If x
1,1∈ D, then the subset D
0= D − {x
1,1} ∪ {x
1,2, x
2,1} is also a dominating set of P
m¤P
n. Moreover, since N
Pm¤Pn(x
1,1) = {x
1,2, x
2,1} ⊂ D
0, then x
1,1is an isolated vertex of the induced subgraph hV (P
m¤P
n) − D
0i
Pm¤Pnof a graph P
m¤P
n. It means that D
0is a split dominating set of P
m¤P
n, with |D
0| ≤ γ(P
m¤P
n) + 1.
If x
1,1∈ D, then it must be that x /
1,2∈ D or x
2,1∈ D (otherwise x
1,1would not be dominated by D in P
m¤P
n). Assume that x
1,2∈ D, then D
0= D ∪ {x
1,2} is a split dominating set of P
m¤P
nand |D
0| ≤ |D| + 1 = γ(P
m¤P
n) + 1, as desired.
Thus γ
s(P
m¤P
n) ≤ γ(P
m¤P
n) + 1, for any m, n ≥ 2 and the proof is complete.
In [2] it was obtained that lim
n,m→∞γ(Pmmn¤Pn)=
15. As a consequence from
the above fact and from Theorem 1 we obtain the following
Corollary 2.
n,m→∞
lim
γ
s(P
m¤P
n)
mn = 1
5 . The following result was proved in [2].
Theorem 3 [2]. For n ≥ 2,
γ(P
2¤P
n) =
» n + 1 2
¼ .
Inspired by this result we shall calculate the split domination number of P
2¤P
n, for n ≥ 2. Before proceeding we give a few necessary results.
Let V (P
2) = {v
1, v
2} and V (P
n) = {u
1, u
2, . . . , u
n}. For convenience, in the rest of the paper we will write x
iinstead of (v
1, u
i) ∈ V (P
2¤P
n) and y
iinstead of (v
2, u
i) ∈ V (P
2¤P
n), for i = 1, 2, . . . , n. Hence V (P
2¤P
n) = {x
i, y
i: i = 1, 2, . . . , n} and E(P
2¤P
n) = {x
ix
i+1, y
iy
i+1, x
iy
i, x
ny
n: i = 1, 2, . . . , n − 1}.
Lemma 4. If n ≡ 2(mod 4), n ≥ 2, then
D = {x
i: i ≡ 1(mod 4)} ∪ {y
j: j ≡ 3(mod 4)} ∪ {y
n} is the γ
s(P
2¤P
n)-set with |D| = d
n+12e.
P roof. Let D = {x
i: i ≡ 1(mod 4)} ∪ {y
j: j ≡ 3(mod 4)} ∪ {y
n} be a subset of V (P
2¤P
n).
We show that any vertex of P
2¤P
nis either in D or it is adjacent to some vertex from D. Let r be an integer not greater than n.
If r = 4q, q ≥ 1, then the vertex x
ris adjacent to x
r+1= x
4q+1∈ D and y
ris adjacent to y
r−1= y
4q−1∈ D.
If r = 4q + 1, q ≥ 0, then x
r∈ D and y
ris adjacent to x
r.
If r = 4q + 2, q ≥ 0, then x
ris adjacent to x
r−1∈ D. If r = n, then y
r= y
n∈ D and if r < n, then y
ris adjacent to y
r+1∈ D.
Finally, if r = 4q + 3, q ≥ 0, then y
r∈ D and x
ris adjacent to y
r. All this together gives that D is a dominating set of P
2¤P
n.
Let n = 4s + 2, s ≥ 0. We state that |D| = d
n+12e. Indeed, partition
V (P
2¤P
n) into subsets B
i= {x
4i−3, y
4i−3, . . . , x
4i, y
4i}, for i = 1, 2, . . . , s
and B
s+1= {x
n−1, y
n−1, x
n, y
n}. Note that |D ∩ B
i| = 2, for i = 1, 2, . . . , s + 1. Thus |D| = 2s + 2 = d
n+12e = γ(P
2¤P
n), by Theorem 3. Since N
P2¤Pn(x
n) = {x
n−1, y
n} ⊂ D, hence x
nis an isolated vertex of hV (P
2¤P
n) − Di
P2¤Pn. Thus this induced subgraph is disconnected. All this together gives that D is a γ
s(P
2¤P
n)-set, since D is a split dominating set of P
2¤P
nwith the minimum cardinality. Hence the result is true.
Lemma 5. If n ≡ 0(mod 4), n ≥ 2, then
D = {x
i: i ≡ 1(mod 4)} ∪ {y
j: j ≡ 3(mod 4)} ∪ {x
n} is the γ
s(P
2¤P
n)-set with |D| = d
n+12e.
P roof. Let D be as in the statement of the theorem. Arguing similarly as in the proof of above lemma, it follows that D is a dominating set of P
2¤P
n. Now, we show that |D| = d
n+12e. Put n = 4s and partition V (P
2¤P
n) into the subsets B
i= {x
4i−3, y
4i−3, . . . , x
4i, y
4i}, for i = 1, 2, . . . , s. It is easy to observe that |D ∩ B
i| = 2, for i = 1, 2, . . . , s − 1 and |D ∩ B
s| = 3. Hence
|D| = 2(s−1)+3 = 2s+1 = d
n+12e = γ(P
2¤P
n), as desired. Finally, observe that y
nis an isolated vertex of hV (P
2¤P
n) − Di
P2¤Pn. This means that the last subgraph is disconnected and as a consequence D is a split dominating set of P
2¤P
n. Since D is also a γ(P
2¤P
n)-set, it is a γ
s(P
2¤P
n)-set, as required.
Lemma 6. Let n ≥ 5 be odd and let D be a γ(P
2¤P
n)-set. Then exactly one of x
1and y
1belong to D.
P roof. Let n = 2k + 1 with k ≥ 2 and let D be a γ(P
2¤P
n)-set. Assume that x
1, y
1∈ D, then it must be that x /
2, y
2∈ D (otherwise x
1or y
1would not be dominated by D). Since n ≥ 5 is odd, then {x
3, y
3} ⊂ V (P
2¤P
n).
Moreover x
3, y
3∈ D. Indeed, without loss of generality, suppose that x /
3∈ D. Then D ∪ {y
1} − {x
2, y
2} is a dominating set of P
2¤P
n, having the cardi- nality |D| − 1. This contradicts the fact that D is the minimum dominating set of P
2¤P
n.
So, we have x
1, y
1, x
3, y
3∈ D and x /
2, y
2∈ D. Consider two induced subgraphs of P
2¤P
n:
X
1= h{x
1, y
1, x
2, y
2, x
3, y
3}i
P2¤Pnand
X
2= h{x
4, y
4, . . . , x
n, y
n}i
P2¤Pn.
Since X
2∼ = P
2¤P
n−3, then by Theorem 3 we have γ(X
2) = d
n−22e = d
2k−12e = k. Further |D| = γ(X
1) + γ(X
2) = 2 + k = d
n+32e > d
n+12e = γ(P
2¤P
n), — a contradiction, since D is a γ(P
2¤P
n)-set.
Now, assume that x
1and y
1∈ D, then x
2, y
2, x
3, y
3∈ D (otherwise / there would exist a dominating set of P
2¤P
nwith order strictly less than the cardinality of D). Arguing as above, for X
1= h{x
1, y
1, x
2, y
2}i
P2¤Pnand X
2= h{x
3, y
3, . . . , x
n, y
n}i
P2¤Pn, we also come to a contradiction. Hence the proof is complete.
In [2] the following was proved
Lemma 7 [2]. If n ≥ 5 and n is odd, then
D = {x
i: i ≡ 1(mod 4)} ∪ {y
j: j ≡ 3(mod 4)}
is the γ(P
2¤P
n)-set with |D| = d
n+12e.
Lemma 8. Let n ≥ 5 be odd and let D be a γ(P
2¤P
n)-set. Then
|D ∩ {x
i, y
i, x
i+1, y
i+1}| = 1, for i = 1, 2, . . . , n − 1.
P roof. We prove this lemma by induction. First consider the base case, when n = 5. By Lemma 6, either x
1∈ D or y
1∈ D and x
5∈ D or y
5∈ D.
Since γ(P
2¤P
5) = 3, then
|D ∩ {x
2, y
2, x
3, y
3, x
4, y
4}| = 1.
If x
3, y
3∈ D, then x /
3or y
3is not dominated by D in P
2¤P
5. So it must be that either x
3∈ D or y
3∈ D. Thus the result holds for n = 5.
Assume that the result holds for n = 2k + 1 and consider n = 2k + 3. By Lemma 6, either x
1∈ D or y
1∈ D. If x
2, y
2∈ D, then by the assumption /
|D ∩ {x
i, y
i, x
i+1, y
i+1}| = 1, for i = 3, 4, . . . , n − 1. Moreover,
|D ∩ {x
1, y
1, x
2, y
2}| = 1 and
|D ∩ {x
2, y
2, x
3, y
3}| = 1.
Thus the result holds for n = 2k + 3.
If x
2∈ D or y
2∈ D, then D
1= D ∩ {x
i, y
i: i = 4, . . . , n} is a γ(P
2¤P
2k)-set and |D
1| = d
2k+12e = k + 1, by Theorem 3. Thus
|D| ≥ |D
1| + 2 = k + 3 >
» 2k + 3 2
¼
= γ(P
2¤P
2k+3) but this is impossible, since D is a γ(P
2¤P
2k+3)-set.
Hence the result is true for all odd n ≥ 5.
Theorem 9. For n ≥ 2,
γ
s(P
2¤P
n) =
( §
n+12
¨ , if n is even or n = 3,
§
n+12