885
Views
3
CrossRef citations to date
0
Altmetric
Articles

Domination parameters of generalized Sierpiński graphs

, &
Pages 4-10 | Received 30 Mar 2022, Accepted 17 Oct 2022, Published online: 02 Nov 2022

Abstract

In this paper, we obtain the Italian domination number, perfect Italian domination number and double Roman domination number of generalized Sierpiński graph S(G,2), where G is a cycle Cn, n4, a complete bipartite graph K1,q or K2,q,q2 and a bistar Bm,n,m,n3.

AMS Subject Classification:

1. Introduction

Let G be a simple graph with vertex set V(G) and edge set E(G). If there is no ambiguity in the choice of G, then we write V(G) and E(G) as V and E, respectively. The number of vertices and edges of the graph G are denoted by n(G) and m(G), respectively. The open neighbourhood of a vertex vV is the set N(v)={u:uvE} and the vertices in N(v) are called the neighbours of v. |N(v)| is called the degree of the vertex v in G and is denoted by dG(v), or simply d(v). The vertices with degree one are called leaves, whereas the vertices with degree n1 are called universal vertices. Let Cn denote a cycle on n vertices and Kp,q denote a complete bipartite graph on p + q vertices, where one partition contains p vertices and other partition contains q vertices. The bistar Bm,n is the graph obtained by joining the center vertices of K1,m and K1,n by an edge.

Let f be a function defined on the vertex set of a graph G. The weight of f is defined as f(V)=vVf(v). A Roman dominating function on G is a function f:V(G){0,1,2} such that every vertex uV with f(u)=0 has at least a neighbor vNG(u) satisfying f(v)=2. The Roman domination number of G, denoted by γR(G), is the minimum weight among all Roman dominating functions on G. Cockayne, Dreyer, S. M. Hedetniemi and S. T. Hedetniemi [Citation14] introduced the concept of Roman Domination in graphs, and since then a lot of related variations and generalizations have been studied (see [Citation7, Citation11–13]).

An Italian dominating function – IDF (perfect Italian dominating function – PID-function) of a graph G is a function f:V(G){0,1,2} satisfying the condition that for every vV with f(v)=0, we have uN(v)f(u)2 (uN(v)f(u)=2), i.e., either v is adjacent to at least one vertex u with f(u)=2 or at least two vertices x and y with f(x)=f(y)=1 (i.e., all the neighbours of v are assigned the weight 0 by f except for exactly one vertex u for which f(u)=2 or for exactly two vertices u and w for which f(u)=f(w)=1). The Italian domination number, γI(G) (perfect Italian domination number, γIp(G)) is the minimum weight of an Italian dominating function [Citation24] (perfect Italian dominating function [Citation23]). The Italian dominating function with weight γI(G) is called a γI-function [Citation10]. The sum of the weights of the vertices of H is denoted by f(H), where H is any subgraph of G. i.e., f(H)=uV(H)f(u). The study of Italian domination was initiated by Chellai et al. in [Citation10] and they called Italian domination as Roman {2}-domination. Italian domination is fairly a new concept and there are only a handful of papers on Italian domination. Interested readers may refer to [Citation2, Citation17, Citation18, Citation20, Citation22, Citation24, Citation26, Citation28] and [Citation35]. In [Citation29], the exact value of perfect Italian domination number for Cartesian product of some special graphs is obtained. A relation between the Roman domination number and the perfect Italian domination number of a graph G is obtained and the corresponding realization problem is also solved. In [Citation32], the authors characterize the graphs G with γIp(G) equal to 2 and 3 and determined the exact value of the parameter for several simple-structured graphs. It is also proved that it is NP-complete to decide whether a given bipartite graph admits a perfect Italian dominating function of weight k.

Given a graph G=(V,E), a function f:V{0,1,2,3} having the property that if f(v) = 0, then there exist v1,v2N(v) such that f(v1)=f(v2)=2 or there exists wN(v) such that f(w) = 3, and if f(v) = 1, then there exists wN(v) such that f(w)2 is called a double Roman dominating function (DRDF). The double Roman domination number, γdR(G), is the minimum among the weights of DRDFs on G. If f:V{0,1,2,3} is a function defined on V(G), then {vV(G):f(v)=i} is denoted by Vif. (If there is no ambiguity, Vif is written as Vi [Citation9]. The study of double Roman domination was initiated by R. A. Beeler, T. W. Haynes and S. T. Hedetniemi in [Citation9]. Interested readers may refer [Citation3–6, Citation21, Citation34, Citation36] and [Citation38]. Domination parameters in some classes of graphs have been studied by several authors (see [Citation15, Citation16]).

Let G=(V,E) be a non-empty graph of order n2, and t a positive integer. Let Vt be the set of words of length t on alphabet V. A word u of length t is denoted by u1u2ut. The graph S(Kn,t),t1, was introduced by Klavžar and Milutinović in [Citation30] and was later called as Sierpiński graphs in [Citation31]. S(Kn,t) has vertex set Vt and {u,v} is an edge if and only if there exists i{1,2,,t} such that: (i) uj=vj, if j<i; (ii) uivi; (iii) uj=vi and vj=ui, if j>i. The vertices of the form uuuu are called extreme vertices of S(Kn,t).

The generalized Sierpiński graph of a graph G, denoted by S(G, t), is a graph with vertex set Vt and edge set {{wuiujr1,wujuir1}:{ui,uj}E,ij;r{1,2,,t};wVtr} [Citation19]. Note that S(G,1) is G itself. If V={1,2,,n} is the vertex set of G, then in S(G,2) Vi={ij:j=1,2,,n} induces a copy of G for each i{1,2,,n}. The subgraph induced by Vi is denoted by Gi, for i{1,2,,n}. gives S(C5,1) and S(C5,2).

Figure 1. Sierpiński graphs S(C5,t),t=1,2.

Figure 1. Sierpiński graphs S(C5,t),t=1,2.

The value of various domination parameters of Sierpiński graphs were studied in [Citation6, Citation27, Citation33] and [Citation37]. The reader may refer to the survey paper [Citation25]. For any graph theoretic terminology and notations not mentioned here, the readers may refer to [Citation8].

The following results are useful in this paper.

Theorem 1.1.

[Citation10] For a cycle Cn and a path Pn, γI(Cn)=n2,γI(Pn)=n+12.

Theorem 1.2.

[Citation32] For a cycle Cn, γIp(Cn)=n2.

Proposition 1.3.

[Citation9] In a double Roman dominating function of weight γdR(G), no vertex needs to be assigned the value 1.

Proposition 1.4.

[Citation1] For n3, γdR(Cn)={nif  n0,2,3,4  (mod6),n+1  if  n1,5  (mod6).

Proposition 1.5.

[Citation1] For n1, γdR(Pn)={nif  n0  (mod3),n+1  if  n1or2(mod3).

In this paper, we obtain the exact values of the Italian domination number, the perfect Italian domination number and the double Roman domination number of the generalized Sierpiński graph S(G,2), where G is a cycle Cn, n4, a complete bipartite graph K1,q or K2,q,q2 and a bistar Bm,n,m,n3.

2. Main Results

For n = 3, CnK3 and γI(S(Kn,2)),γIp(S(Kn,2)) and γdR(S(Kn,2)) were discussed in [Citation27, Citation28] and [Citation6], respectively. Hence, in this section, we consider Cn, for n4 only.

Theorem 2.1.

The Italian domination number of the generalized Sierpiński graph S(Cn,2) is γI(S(Cn,2))=nn2, for n4.

Proof.

Let V(Cn)={v1,v2,,vn}. Then S(Cn,2) has the vertex set {vivj:i,j{1,2,,n}} and edge set {(vivj,vivk):vjvkE(Cn)}{(vivj,vjvi):vivjE(Cn)}. In S(Cn,2) there are n copies of Cn and we know that γI(Cn)=n2. Therefore, if we take a γI-function in each copy of Cn, we get an IDF of S(Cn,2), so that γI(S(Cn,2))nn2.

For the reverse inequality, note that in Cni,vivi is the extreme vertex and vivi1 and vivi+1 are the vertices which are adjacent to Cni1 and Cni+1, respectively. The vertices other than vivi1,vivi,vivi+1 form a path on n3 vertices, say Pn3i. Therefore, f(Pn3i{vivi2,vivi+2})n5+12=n42. To Italian dominate vivi2,vivi,vivi+2 (which do not have any adjacency outside Cni) we need minimum weight 2. Therefore, f(Cni)n42+2=n2 so that γI(S(Cn,2))nn2. Hence, γI(S(Cn,2))=nn2.

Corollary 2.2.

The perfect Italian domination number of the generalized Sierpiński graph S(Cn,2) is γIp(S(Cn,2))=nn2, for n4.

Proof.

We know that γIp(Cn)=n2. The proof is similar to that of Theorem 2.1. □

Lemma 2.3.

If Pn is a path on n vertices, where n is a multiple of 3, then the weight of a γdR-function assigned to the end vertices is always zero.

Proof.

Let Pn be the path v1v2vn. Let f be any γdR-function with f(v1)=1. Then v1 cannot double Roman dominate any other vertex. Hence, f(Pnv1)n1+1=n, so that f(Pn)n+1 which is a contradiction.

Now let f be a γdR-function with f(v1)=2. If f(v2)1, then none of the vertices of Pn{v1,v2} are double Roman dominated by vertices outside Pn{v1,v2}. Consequently, f(Pn{v1,v2})n2+1=n1, so that f(Pn)n1+2=n+1, which is a contradiction. If f(v2)>1 then we can find a DRDF g as follows. g(v1)=0,g(v2)=3 and g(vi)=f(vi) for i=3,4,,n. Clearly, g is a DRDF with weight less than that of f, which is also a contradiction. Hence, we conclude that f(v1)2.

Similarly, the case when f(v1)=3 also leads to a contradiction and so f(v1)=0. In the same manner, we can also prove that f(vn)=0.

Lemma 2.4.

If Cn is a cycle, where n is an odd multiple of 3 then in any γdR-function, no vertex is assigned the weight 2.

Proof.

Let Cn be a cycle v1v2vn and let f be a γdR-function with at least one vertex assigned the weight 2. For definiteness, let f(v1)=2. If f(v2) is non-zero, then f(v1)+f(v2)4 and f(Cn{v1,v2,v3,vn})n4+1=n3, so that f(Cn)n3+4=n+1 which is a contradiction. If f(v2)=0, then f(v3) must be non-zero. If f(v3)=3 then f(v1)+f(v3)=5 and f(Cn{v1,v2,v3,v4,vn})n5+1=n4 so that f(Cn)n4+5=n+1 which is also a contradiction. Hence, f(v3)=2. Continuing this argument, we get f(vi)=0 or 2 according as i is even or odd, respectively, so that f(Cn)=n+1 which is a contradiction. Hence, the result follows. □

Theorem 2.5.

The double Roman domination number of the generalized Sierpiński graph S(Cn,2) is γdR(S(Cn,2))={n2n2ifn=3k,k1iseven,n2n3ifn=3k,k1isodd,n(n1)ifn=3k+1,k1,n2ifn=3k+2,k1.

Proof.

Let V(Cn)={v1,v2,,vn}. Then S(Cn,2) has the vertex set {vivj:i,j{1,2,,n}} and edge set {(vivj,vivk):vjvkE(Cn)}{(vivj,vjvi):vivjE(Cn)}. We consider the following cases.

Case 1: n=3k,k1 is even.

Define the following function on S(Cn,2). f(vivj)={3ifi=2,4,,n;j=i+3l(modn);l=1,2,,k1,2ifi=1,3,,n1;j=i+1+2l(modn),l0andi=j=2,4,,n,0otherwise. Clearly, f is a DRDF and f(V)=3n2(k1)+2n2n2+2n2=n2n2. Hence, γdR(S(Cn,2))n2n2.

For the reverse inequality, note that in each Cni,vivi is the extreme vertex and vivi1 and vivi+1 are the only vertices adjacent to other copies of Cn. The remaining vertices (other than vivi1,vivi,vivi+1) form a path on n3 vertices, say Pn3i. Let f be any DRDF of S(Cn,2). If f(vivi1)+f(vivi+1)=0, then f(vivi)2 and f(Pn3i)n3, so that f(Cni)n3+2=n1. If f(vivi1)+f(vivi+1)=2, then one of f(vivi1) and f(vivi+1) must be 2. For definiteness, let f(vivi+1)=2. Then f(Pn3ivivi+2)n4+1=n3. In this case, f(vivi)2 so that f(Cni)n3+2+2=n+1. If f(vivi1)+f(vivi+1)=3, then one of f(vivi1) and f(vivi+1) is 3. For definiteness, let f(vivi+1)=3. Then f(Pn3ivivi+2)n4+1=n3 so that f(Cni)n. If f(vivi1)+f(vivi+1)4 then f(Pn3ivivi2,vivi+2)n5+1=n4 so that f(Cni)n. Thus, in all cases, f(Cni)n1 and if f(Cni)=n1 then f(Pn3i)=n3,f(vivi1)=f(vivi+1)=0 and f(vivi)=2. We claim that if f(Cni0)=n1 then f(Cni0+1)n. Since f(Pn3i0)=n3, then f(viovi02)=f(vi0vi0+2)=0 by Lemma 2.3. Therefore, since f(viovi01)=f(vi0vi0+1)=0, to double Roman dominate vi0vi0+1,f(vi0+1vi0)2 so that f(Cni0+1)n. Thus if f(Cni0)=n1 then f(Cni0+1)n so that f(S(Cn,2))nn2+(n1)n2=n2n2.

Case 2: n=3k, k1 is odd.

As in case 1, for any γdR-function f of S(Cn,2) f(Cni)n1 for each i=1,2,,n. We claim that if f(Cni0)=n1, then both f(Cni01) and f(Cni0+1) is at least n. If f(Cni0)=n1, then f(vi0vi0)=2 and f(viovi01)=f(vi0vi0+1)=0. So to double Roman dominate vi0vi0+1,f(vi0+1vi0) must be at least 2. If f(vi0+1vi0)=2, then f(Cni0+1)n+1 by Lemma 2.4 and so f(vi0+1vi0)=3. If f(vi0+1vi0)=3 then f(vi0+1vi0+2) must be zero in order to make f(Cni0+1)=n. i.e., the only vertex which is outside Cni0+1 and double Roman dominated by vertices in Cni0+1 is vi0vi0+1. Similarly, f(vi01vi0)=3,f(Cni01)=n and the only vertex which is outside Cni01 and double Roman dominated by vertices in Cni01 is vi0vi01. Thus, for three consecutive copies of Cn, at most one copy can have weight n1 and hence f(S(Cn,2))n2n3.

Case 3: n=3k+1,k1.

Define the following function on S(Cn,2) f(vivj)={3ifi{1,2,,n}andj=i+1+3l(modn);l=0,1,,k1,0otherwise. Clearly, f is a DRDF and f(V)=n3k=n(n1). Hence, γdR(S(Cn,2))n(n1).

For the reverse inequality, note that f(Cni)n1 for all i{1,2,,n} as in above cases and hence f(S(Cn,2))n(n1).

Case 4: n=3k+2,k1.

Define the following function on S(Cn,2). f(vivj)={3ifi{1,2,,n},j=i+1+3l(modn),l=0,1,,k1,2ifi{1,2,,n},j=i3(modn),0otherwise. Clearly, f is a DRDF and f(V)=n(3k+2)=n2. Hence, γdR(S(Cn,2))n2.

For the reverse inequality, the proof is similar to case 3 except when f(vivi1)+f(vivi+1)4. If f(vivi1)+f(vivi+1)=4, then f(vivi1)=f(vivi+1)=2 and hence Pn3i{vivi1,vivi+1} can be considered as a path on n1 vertices of which none of the vertices are double Roman dominated by vertices from copies of Cn other than Cni. Consequently, f(Cni)f(Pn3i{vivi1,vivi+1})n1+1=n. If f(vivi1)+f(vivi+1)5, then f(Pn3i{vivi2,vivi+2})n5 and so f(Cni)n. Thus in all cases, f(Cni)n and hence f(S(Cn,2))n2.

When p = 1 and q = 1 Kp,q is K2 and it is proved in [Citation28] that γI(S(Kn,2))=2n1. Therefore, γI(S(K1,1,2))=3 and also γIp(S(K1,1,2))=3. Also in [Citation6], it is found that γdR(S(Kn,2))=3n1 and hence γdR(S(K1,1,2))=5. When p = 1 and q = 2 then Kp,q=P3 and it is a simple observation that γI(S(P3,2))=5,γIp(S(P3,2))=6 and γdR(S(P3,2))=8. Hence, in the following theorems, we consider the case p = 1 and q3.

Theorem 2.6.

The Italian domination number of the generalized Sierpiński graph S(K1,q,2) is γI(S(K1,q,2))=2q+1,q3.

Proof.

Let v1,v2,,vq+1 be the vertices of K1,q where v1 is the universal vertex. Define the following function on S(K1,q,2) as follows. f(v)={2ifv=viv1,i=2,3,,q+1,1ifv=v1v1,0otherwise. Clearly, f is an IDF and f(V)=2q+1 and hence, γI(S(K1,q,2))2q+1.

Since q3, each copy of K1,qi for i=2,3,,q+1, contains three or more leaves. So to Italian dominate these vertices we need to assign weight at least 2 to the vertex viv1. Now in K1,q1, the vertices v1vj,j=2,3,,q+1 are Italian dominated by vjv1. So the only vertex which is not Italian dominated in K1,q1 is v1v1. Hence, we have to assign weight 1 to v1v1, so that, γI(S(K1,q,2))2q+1. Therefore, γI(S(K1,q,2))=2q+1.

Theorem 2.7.

The perfect Italian domination number of the generalized Sierpiński graph S(K1,q,2) is γIp(S(K1,q,2))=2(q+1),q3.

Proof.

Let v1,v2,,vq+1 be the vertices of K1,q where v1 is the universal vertex. Define the following function on S(K1,q,2) as follows. f(v)={2ifv=viv1,i=2,3,,q+1,andv=v1v2,0otherwise. Clearly, f is a PID function and f(V)=2q+2=2(q+1). Therefore, γIp(S(K1,q,2))2(q+1).

In each K1,qi for i=2,3,,q+1, there are q vertices which are not adjacent to vertices of other copies of K1,q. Hence, for any γIp-function f of S(K1,q,2),f(K1,qi)2,i=2,3,,q+1. Also, it is optimum to assign the weight 2 to viv1 where i=2,3,,q+1. Now, the only vertex which is not perfect Italian dominated is v1v1. We cannot give the weight 1 to v1v1, since, in this case uN(v1vi)f(u)=3, for i=2,3,,q+1. Hence, we have to give weight to the vertices of K1,q1, so that uN(v1v1)f(u)=2, which results in f(K1,q1)=2. Therefore, γIp(S(K1,q,2))2q+2=2(q+1). Hence, γIp(S(K1,q,2))=2(q+1).

Theorem 2.8.

The double Roman domination number of the generalized Sierpiński graph S(K1,q,2) is γdR(S(K1,q,2))=3q+2,q3.

Proof.

The proof is similar to that of Theorem 2.6 with the difference that the weights 2 and 1 are replaced by 3 and 2, respectively. □

When p=q=2,Kp,q=C4 and it is discussed in the section 2.

Theorem 2.9.

The Italian domination number of the generalized Sierpiński graph S(K2,q,2) is γI(S(K2,q,2))=2(q+2) for p = 2, q3.

Proof.

Let V(K2,q)={v1,v2,,vq+2} be the vertex set of K2,q and d(v1)=d(v2)=q. Define the following function on S(K2,q,2) as follows. f(v)={1ifv=viv1,viv2,i{1,2,,q+2},0otherwise. It can be easily verified that f is an IDF and f(V)=2(q+2). Therefore, γI(S(K2,q,2))2(q+2).

In each K2,qi there are at least two vertices which are not adjacent to vertices of other copies of K2,q. So to Italian dominate K2,qi we need at least weight 2. Therefore, γI(S(K2,q,2))2(q+2). Hence, γI(S(K2,q,2))=2(q+2).

Proposition 2.10.

The perfect Italian domination number of generalized Sierpiński graph S(K2,3,2) is γIp(S(K2,3,2))=11.

Proof.

Let {v1,v2,v3,v4,v5} be the vertex set of K2,3 where {v1,v2} and {v3,v4,v5} are the partite sets of the vertex set. Define a function f on S(K2,3,2) as follows. f(v)={2ifv=v2v5,v5v1,1ifvivj,i=3,4,j=1,2,andv=v1v3,v1v4,v2v2,0otherwise. It can be easily verified that f is a PID- function and f(V)=11 so that γIp(S(K2,3,2))11. (This labeling is illustrated in .)

Figure 2. Illustration of S(K2,3,2).

Figure 2. Illustration of S(K2,3,2).

In each copy of K2,3 there are at least 2 vertices which are not adjacent to the vertices of other copies of K2,3. Hence to perfect Italian dominate each K2,3i we need minimum weight 2. If possible, assume that f(K2,3i)=2 for every i=1,2,3,4,5. In particular, f(K2,31)=2. Since v1v1 and v1v2 are not adjacent to any vertices of other copies of K2,3, these vertices must be perfect Italian dominated by the vertices of K2,31. Then, one of the following cases may arise.

Case 1: f(v1v1)=f(v1v2)=1.

Since f(K2,31)=2,f(v1vj)=0. Since f is a PID function uN(v1vj)f(v1vj)=2 and hence f(vjv1)=0 for j=3,4,5. Now, to perfect Italian dominate vjv1, we must have k=35f(vjvk)=2,j=3,4,5. This implies, there exists at least one vertex vjvk with f(vjvk)=0. To perfect Italian dominate this vjvk, we must have f(vjv2)=2, which is a contradiction to the fact that f(K2,3j)=2.

Case 2: j=35f(v1vj)=2.

Then for the vertices v1v3,v1v4 and v1v5 either two vertices have weight 1 or one vertex has weight 2.

Subcase(a): Two vertices have weight 1 and the third vertex has weight 0.

For definiteness, let f(v1v3)=f(v1v4)=1 and f(v1v5)=0. Now, to perfect Italian dominate v1v5 we must have f(v5v1)=2. Since f(K2,35)=2,f(v5v2)=0. To perfect Italian dominate v5v2 we must have f(v2v5)=2. Since f(K2,32)=2,f(v2v3)=f(v2v4)=0. To perfect Italian dominate v2v3 and v2v4 we must have f(v3v2)=f(v4v2)=2. But the vertices v3v1 and v4v1 are not perfect Italian dominated. To perfect Italian dominate these vertices we need more weight, which contradicts the fact that f(K2,3i)=2 for all i.

Subcase(b): One vertex has weight 2 and other vertices have weight 0.

For definiteness, let f(v1v3)=2, and f(v1vi)=0, for i=1,2,4,5. To perfect Italian dominate v1v4 and v1v5 we must have f(v4v1)=f(v5v1)=2. Since f(K2,34)=f(K2,35)=2, we have f(v4v2)=f(v5v2)=0. To perfect Italian dominate v4v2 and v5v2 we must have f(v2v4)=f(v2v5)=2, which is a contradiction to the fact that f(K2,32)=2.

Therefore, γIp(S(K2,3,2))11. Hence, γIp(S(K2,32))=11.

Theorem 2.11.

The perfect Italian domination number of the generalized Sierpiński graph S(K2,q,2) is γIp(S(K2,q,2))=2(q+3), for p=2,q4.

Proof.

Let V(K2,q)={v1,v2,,vq+2} be the vertex set of K2,q where {v1,v2} and {v3,v4,,vq+2} are the partite sets of K2,q. Define a function on S(K2,q,2) as follows. f(v)={2ifv=viv3,i=1,2,1ifv=viv1,viv2,i=3,4,,q+2,v=viv1,i=1,2,0otherwise. Clearly, f is a PID-function and f(V)=2(q+3) so that γIp(S(K2,q,2))2(q+3).

Since in each copy of K2,qi there are at least two vertices which are not adjacent to vertices of other copies of K2,q, we need at least weight 2 to perfect Italian dominate each K2,qi. If there are two copies of K2,q with f(K2,qi)=3, then f(S(K2,q,2))=2(q+3). Therefore, if possible assume that, there exists exactly one copy of K2,q with f(K2,qi)=3 and all other copies have weight 2. Then either f(K2,q1)=2 or f(K2,q2)=2. For definiteness, let f(K2,q1)=2. Since both v1v1 and v1v2 are perfect Italian dominated by the vertices of K2,q1, one of the following cases arises.

Case 1: f(v1v1)=f(v1v2)=1.

Since f(K2,q1)=2,f(v1vj)=0 for j=3,4,,q+2. Since f is a perfect Italian dominating function uN(v1vj)f(u)=2 so that f(vjv1)=0 for j=3,4,,q+2. To perfect Italian dominate vjv1, we must have uN(vjv1)f(u)=2. This implies, there exists at least two vertices vjvk such that f(vjvk)=0 for j,k=3,4,,q+2, and hence f(vjv2)=0, which implies f(K2,qi)=4 for i=3,4,,q+2, which is a contradiction.

Case 2: j=3q+2f(v1vj)=2.

Then for the vertices v1v3,v1v4,,v1vq+2 either two vertices have weight 1 or one vertex has weight 2.

Subcase (a): f(v1vj)=1 for exactly two js, and f(v1vj)=0 for all other js,j=3,4,,q+2.

For definiteness, let f(v1v3)=f(v1v4)=1. Since f(K2,q1)=2,f(v1vj)=0 for j=5,6,,q+2. To perfect Italian dominate v1vj for j=5,6,,q+2 we must have f(vjv1)=2. Since f(K2,qj)=2,f(vjv2)=0. To perfect Italian dominate vjv2 we must have f(v2vj)=2, for j=5,6,,q+2. This implies, f(K2,q)>4, which is a contradiction.

Subcase (b): f(v1vj)=2 for exactly one j and f(v1vj)=0 for all other js,j=3,4,,q+2.

For definiteness, let f(v1v3)=2 and f(v1vj)=0 for j=4,5,,q+2. To perfect Italian dominate v1vj we must have f(vjv1)=2. Since f(K2,qj)=2 we must have f(vjv2)=0. To perfect Italian dominate vjv2 we must have f(v2vj)=2. This implies that f(K2,q2)>6, which is a contradiction.

Therefore, γIp(S(K2,q,2))2q+6. Hence, γIp(S(K2,q,2))=2(q+3).

Theorem 2.12.

The double Roman domination number of generalized Sierpiński graph S(K2,3,2) is γdR(S(K2,3,2))=18.

Proof.

Let {v1,v2,v3,v4,v5} be the vertex set of K2,3 where {v1,v2} and {v3,v4,v5} are the partite sets of the vertex set. Define a function f as follows. f(v)={3ifv=v1v3,v2v4,v2v5,v3v2,v4v1,v5v1,0otherwise. Clearly, f is a DRDF and γdR(S(K2,3,2))18.

To prove the reverse inequality, note that in each copy of K2,3 in S(K2,3,2) there are at least two vertices which are not adjacent to other copies of K2,3. Hence, for any γdR-function f of S(K2,3,2),f(K2,3i)3 for all i{1,2,,5} so that f(S(K2,3,2))15. Therefore, in order to prove γdR(S(K2,3,2))=18, we need only to prove that there does not exist any DRDF with weight 15, 16 or 17.

Claim 1: There does not exist a DRDF f with f(S(K2,3,2))=15.

Proof of Claim 1:

If there exists a DRDF f with f(S(K2,3,2))=15, then f(v1vj)=3 for exactly one j=3,4,5. For definiteness, let f(v1v3)=3. Now to double Roman dominate v1v4 and v1v5,f(v4v1) and f(v5v1) must be three. Since, f(K2,34)=f(K2,35)=3,v4v2 and v5v2 should be double Roman dominated by v2v4 and v2v5 respectively which makes f(K2,32)6 which is a contradiction.

Claim 2: There does not exist a DRDF f with f(S(K2,3,2))=16.

Proof of Claim 2:

If there exists a DRDF f with f(S(K2,3,2))=16, then f(K2,3i0)=4 for exactly one i0 with 1i05 and f(K2,3i)=3, for every i{1,2,,5},ii0. Then at least one among f(K2,31) and f(K2,32) must be 3. For definiteness, let f(K2,31)=3. Since f(K2,3j)4, for every j=2,3,4,5, we get the same contradiction as in the proof of claim 1.

Claim 3: There does not exist a DRDF f with f(S(K2,3,2))=17.

Proof of Claim 3:

If there exists a DRDF f with f(S(K2,3,2))=17, then the following two cases arise.

  1. f(K2,3i0)=5 for exactly one i0 with 1i05 and f(K2,3i)=3, for every i{1,2,,5},ii0.

  2. f(K2,3i)=4 for exactly two is say i0, j0 with 1i0,j05 and f(K2,3i)=3, for every i{1,2,,5},ii0,j0.

In case (i) at least one among f(K2,31) and f(K2,32) must be 3. For definiteness, f(K2,31)=3. Then as in the proof of claim 1 we may take f(v1v3)=f(v4v2)=f(v5v2)=3. If f(K2,34)=f(K2,35)=3, then we get the same contradiction as in the proof of claim 1. Hence one among f(K2,34) and f(K2,35) must be 5. For definiteness let f(K2,34)=5. Since f(K2,35)=3, to double Roman dominate v5v2,f(v2v5)=3. But v2v4 is not get double Roman dominated and hence f(K2,32)3, which is a contradiction. Hence, we conclude that γdR(S(K2,3,2))18.

Now we obtain the Italian domination number, perfect Italian domination number and double Roman domination number of Sierpiński graph of bistar Bm,n, when t = 2.

Theorem 2.13.

The Italian domination number of the generalized Sierpiński graph S(Bm,n,2) is γI(S(Bm,n,2))=4(m+n+1),m,n3.

Proof.

Let V(Bm,n)={v1,v2,,vm,vm+1,vm+2,,vm+n,vm+n+1,vm+n+2} be the vertex set of Bm,n. Let d(vm+n+1)=m+1,d(vm+n+2)=n+1. Define a function f on Bm,n as follows. f(v)={2ifv=vm+n+1vm+n+2,vm+n+2vm+n+1orv=vivm+n+1,vivm+n+2,i{1,2,,m+n},0otherwise. Clearly, f is an IDF and f(V)=4(m+n)+4=4(m+n+1) so that γI(S(Bm,n,2))4(m+n+1).

In each Bm,ni for i{1,2,,m+n} there are m + n leaves. To Italian dominate these copies we need at least weight 4. For that, it is optimum to assign weight 2 to the vertices vivm+n+1 and vivm+n+2. In Bm,nm+n+1 and Bm,nm+n+2 there are only n and m leaves, respectively. To Italian dominate these vertices we need at least weight 2. Therefore, γI(S(Bm,n,2))4(m+n)+2+2=4(m+n+1). Hence, γI(S(Bm,n,2))=4(m+n+1).

Corollary 2.14.

The perfect Italian domination number of the generalized Sierpiński graph S(Bm,n,2) is γIp(S(Bm,n,2))=4(m+n+1).

Proof.

In the proof of the above theorem we have defined an Italian dominating function with the property that every vertex with weight 0 is adjacent to exactly one vertex with weight 2. Therefore, γIp(S(Bm,n,2))4(m+n+1). We know that γI(S(Bm,n,2))γIp(S(Bm,n,2)). Hence, γIp(S(Bn,n,2))=4(m+n+1).

Theorem 2.15.

The double Roman domination number of the generalized Sierpiński graph S(Bm,n,2) is γdR(S(Bm,n,2))=6(m+n+1),m,n3.

Proof.

The proof is similar to that of Theorem 2.13 with the difference that the weight 2 is replaced by 3. □

References

  • Ahangar, H. A., Chellali, M, Sheikholeslami, S. M. (2017). On the double Roman domination in graphs. Discrete Appl. Math. 232: 1–7.
  • Alizadeh, F., Maimani, H. R., Majd, L. P, Parsa, M. R. (2022). Roman {2}-domination in graphs and graph products. Iran. J. Math. Sci. Inform. In Press.
  • Amjadi, J., Nazari-Moghaddam, V., Sheikholeslami, S. M, Volkmann, L. (2018). An upper bound on the double Roman domination number. J. Comb. Optim. 36(6): 1–9.
  • Anu, V, Aparna, L. S. (2018). Double Roman domination number. Discrete Appl. Math. 244: 198–204.
  • Anu, V, Aparna, L. S. (2021). Impact of some graph operators on double Roman domination number. Int. J. Comb. Graph Theory Appl. 6(1): 97–105.
  • Anu, V, Aparna, L. S. (2020). The double Roman domination number of generalized Sierpiński graphs. Discrete Math. Algorithms Appl. 12(4): Article 2050047.
  • Aparna, L. S, Amala, B. (2022). Variations of Roman domination in Kneser graphs. Manuscript.
  • Balakrishnan, R, Ranganathan, K. (1999). A Text Book of Graph Theory. New York: Springer.
  • Beeler, R. A., Haynes, T. W, Hedetniemi, S. T. (2016). Double Roman domination. Discrete Appl. Math. 211(1): 23–29.
  • Chellali, M., Haynes, T. W., Hedetniemi, S. T, McRae, A. A. (2016). Roman {2}-domination. Discrete Appl. Math. 204: 22–28.
  • Chellai, M., Rad, N. J., Sheikholeslami, S. M, Volkmann, L. (2020). Roman domination in graphs. In: Haynes, T.W., Hedetniemi, S.T., Henning, M. A., eds. Topics in Domination in Graphs. Berlin/Heidelberg: Springer, pp. 365–409.
  • Chellali, M., Jafari Rad, N., Sheikholeslami, S. M, Volkmann, L. (2020). Varieties of Roman Domination II. AKCE Int. J. Graphs Comb. 17(3): 966–984.
  • Chellai, M., Rad, N. J., Sheikholeslami, S. M, Volkmann, L. (2021). Varieties of Roman Domination, Structures of Domination in Graphs. In: Haynes, T.W., Hedetniemi, S.T., Henning, M. A., eds. Topics in Domination in Graphs. Berlin/Heidelberg: Springer, pp. 273–307.
  • Cockayne, E. J., Dreyer, P. A., Jr, Hedetniemi, S. M, Hedetniemi, S. T. (2004). Roman domination in graphs. Discrete Math. 278: 1–3.
  • Deepalakshmi, J., Marimuthu, G., Somasundaram, A, Arumugam, S. (In press). Domination parameters of a splitting graph of a graph. Commun. Comb. Optim.
  • Desormeaux, W. J., Haynes, T. W, Henning, M. A. (2018). Domination parameters of a graph and its complement. Discuss. Math. Graph Theory 38(1): 203–215.
  • Gao, H., Xi, C., Li, K., Zhang, Q, Yang, Y. (2019). The Italian domination number of generalized Petersen graphs P(n,3). Mathematics 7(8): 714. Article 714.
  • Gao, H., Xu, T, Yang, Y. (2019). Bagging approach for Italian domination in Cn□Pm. IEEE Access.
  • Gravier, S., Kovše, M, Parreau, A. (2011). Generalized Sierpinski Graphs 1. https://www.semanticscholar.org/paper/Generalized-Sierpinski-graphs-1-Gravier-Kovse/3c21ce24174d1cdf788e2988702c1e81dc686436.
  • Hajibaba, M, Rad, N. J. (2019). On domination, 2-domination and Italian domination numbers. Util. Math. 111: 271–280.
  • Hao, G., Chen, X, Volkmann, L. (2019). Double Roman domination in digraphs. Bull. Malays. Math. Sci. Soc. 42(5): 1907–1920.
  • Haynes, T. W., Henning, M. A, Volkman, L. (2020). Graphs with large Italian domination number. Bull. Malays. Math. Sci. Soc. 43(6): 4273–4287.
  • Haynes, T. W, Henning, M. A. (2019). Perfect Italian domination in trees. Discrete Appl. Math. 260: 164–177.
  • Henning, M. A, Klostermeyer, W. F. (2017). Italian domination in trees. Discrete Appl. Math. 217: 557–564.
  • Hinz, A. M., Klavžar, S, Zemljicˇ, S. S. (2017). A survey and classification of Sierpiński-type graphs. Discrete Appl. Math. 217: 565–600.
  • Jismy, V, Aparna, L. S. (2022). Impact of vertex addition on Italian domination number. Indian J. Discrete Math. 8(1): 11–20.
  • Jismy, V., Anu, V, Aparna, L. S. (2021). Italian domination and perfect Italian domination on Sierpiński graphs. J. Discrete Math. Sci. Cryptogr. 24(7): 1885–1894.
  • Jismy, V, Aparna, L. S. (2021). Italian domination on Mycielskian and Sierpiński graphs. Discrete Math. Algorithms Appl. 13(4): Article 2150037.
  • Jismy, V, Aparna, L. S. (2022). Perfect Italian domination number of graphs. Palest. J. Math. 11(1): 260–270.
  • Klavžar, S, Milutinović, U. (1997). Graphs S(n,k) and a variant of the tower of Hanoi problem. Czechoslov. Math. J. 47(1): 95–104.
  • Klavžar, S., Milutinović, U, Petr, C. (2002). 1-Perfect codes in Sierpiński graphs. Bull. Austral. Math. Soc. 66(3): 369–384.
  • Lauri, J, Mitillos, C. (2020). Perfect Italian domination on planar and regular graphs. Discrete Appl. Math. 285: 676–687.
  • Liu, C. A. (2020). Domination in Sierpiński Graphs S(Kn,t). arXiv:2008.09807v1.
  • Mojdeh, D. A., Masoumi, I, Volkman, L. (2022). Restrained double Roman domination of a graph. RAIRO-Oper. Res. 56(4): 2293–2304.
  • Poureidi, A, Rad, N. J. (2020). On the algorithmic complexity of Roman {2}-domination. Iran. J. Sci. Technol. Trans. Sci. 44(3): 791–799.
  • Poklukar, D. R, Zerovnik, J. (2022). On the double Roman domination in generalized Petersen graphs P(5k,k). Mathematics 10(1): Article 119.
  • Ramezani, F., Rodríguez-Bazan, E. D, Rodríguez-Velázquez, J. A. (2017). On the Roman domination number of generalized Sierpiński graphs. Filomat 31(20): 6515–6528.
  • Yang, H, Zhou, X. (2020). Some properties of double Roman domination. Discrete Dyn. Nat. Soc. 2020: 1–5. Article 6481092.