Citation: Fu-Tao Hu, Xing Wei Wang, Ning Li. Characterization of trees with Roman bondage number 1[J]. AIMS Mathematics, 2020, 5(6): 6183-6188. doi: 10.3934/math.2020397
[1] | Zhibin Du, Ayu Ameliatul Shahilah Ahmad Jamri, Roslan Hasni, Doost Ali Mojdeh . Maximal first Zagreb index of trees with given Roman domination number. AIMS Mathematics, 2022, 7(7): 11801-11812. doi: 10.3934/math.2022658 |
[2] | Abel Cabrera Martínez, Iztok Peterin, Ismael G. Yero . Roman domination in direct product graphs and rooted product graphs. AIMS Mathematics, 2021, 6(10): 11084-11096. doi: 10.3934/math.2021643 |
[3] | Jian Yang, Yuefen Chen, Zhiqiang Li . Some sufficient conditions for a tree to have its weak Roman domination number be equal to its domination number plus 1. AIMS Mathematics, 2023, 8(8): 17702-17718. doi: 10.3934/math.2023904 |
[4] | Zepeng Li . A note on the bounds of Roman domination numbers. AIMS Mathematics, 2021, 6(4): 3940-3946. doi: 10.3934/math.2021234 |
[5] | Rangel Hernández-Ortiz, Luis Pedro Montejano, Juan Alberto Rodríguez-Velázquez . Weak Roman domination in rooted product graphs. AIMS Mathematics, 2021, 6(4): 3641-3653. doi: 10.3934/math.2021217 |
[6] | Bana Al Subaiei, Ahlam AlMulhim, Abolape Deborah Akwu . Vertex-edge perfect Roman domination number. AIMS Mathematics, 2023, 8(9): 21472-21483. doi: 10.3934/math.20231094 |
[7] | Saeed Kosari, Yongsheng Rao, Zehui Shao, Jafar Amjadi, Rana Khoeilar . Complexity of signed total $k$-Roman domination problem in graphs. AIMS Mathematics, 2021, 6(1): 952-961. doi: 10.3934/math.2021057 |
[8] | Mingyu Zhang, Junxia Zhang . On Roman balanced domination of graphs. AIMS Mathematics, 2024, 9(12): 36001-36011. doi: 10.3934/math.20241707 |
[9] | Abel Cabrera-Martínez, Andrea Conchado Peiró, Juan Manuel Rueda-Vázquez . Further results on the total Italian domination number of trees. AIMS Mathematics, 2023, 8(5): 10654-10664. doi: 10.3934/math.2023540 |
[10] | Chang-Xu Zhang, Fu-Tao Hu, Shu-Cheng Yang . On the (total) Roman domination in Latin square graphs. AIMS Mathematics, 2024, 9(1): 594-606. doi: 10.3934/math.2024031 |
For terminology and notation on graph theory not given here, the reader is referred to Xu [14]. Let G=(V,E) be a finite, undirected and simple graph, where V=V(G) is the vertex set and E=E(G) is the edge set of G. For a vertex x∈V(G), let NG(x)={y∈V(G):xy∈E(G)} be the open set of neighbors of x and NG[x]=NG(x)∪{x} be the closed set of neighbors of x, EG(x)={xy∈E(G):y∈NG(x)} and dG(x)=|EG(x)| be the vertex degree of x.
A subset D⊆V is a dominating set of G if every vertex in V−D has at least one neighbor in D. The domination number of G, denoted by γ(G), is the minimum cardinality among all dominating sets of G. The domination is an important and classic notion that has become one of the most widely researched topics in graph theory. A thorough study of domination appears in the books [7,8] by Haynes, Hedetniemi, and Slater. Among various problems related to the domination number, some focus on graph alterations and their effects on the domination number. Here, we are concerned with the removal of edges from a graph. The bondage number of G, denoted by b(G), is the minimum number of edges whose removal from G results in a graph with domination number larger than that of G. The bondage number was introduced by Fink et at. [5] in 1990. The bondage number are an important parameters for measuring the vulnerability and stability of the network domination under link failure. Xu [15] gave a review article on bondage numbers in 2013.
The Roman dominating function (RDF) on G, proposed by Stewart [13], is a function f:V→{0,1,2} such that each vertex x with f(x)=0 is adjacent to at least one vertex y with f(y)=2. For S⊆V let f(S)=∑u∈Sf(u). The value f(V(G)) is called the weight of f, denoted by f(G). The Roman domination number, denoted by γR(G), is defined as the minimum weight of all Roman dominating functions, that is,
γR(G)=min{f(G):f is a Roman dominating function on G}. |
A Roman dominating function f is called to be a minimum Roman dominating function (MRDF) if f(G)=γR(G).
The Roman bondage number, denoted by bR(G), proposed by Rad and Volkmann [9], of a nonempty graph G is the minimum number of edges whose removal from G results in a graph with larger Roman domination number. Precisely speaking, the Roman bondage number
bR(G)=min{|B|:B⊆E(G),γR(G−B)>γR(G)}. |
Roman domination number has been well studied [3,4].
An edge set B⊆E(G) that γR(G−B)>γR(G) is called the Roman bondage set and the minimum one is called the minimum Roman bondage set. In [2], the authors showed that the decision problem for bR(G) is NP-hard even for bipartite graphs. The Roman bondage number has been further studied for example in [1,6,11,10,12].
In 2001, Rad and Volkmann [9] proved that the Roman bondage number for trees is no more than 3. They proposed a problem that is to determine the trees with Roman bondage number 1. In this paper, we characterize trees with Roman bondage number 1.
Let f=(V0,V1,V2) be a Roman dominating function of G where Vi={v∈V(G):f(v)=i} for i=0,1,2. Let u∈V(G) and f(u)=2, the private neighborhood of u with respect to f is defined as the set
PN(u,f,G)={v∈NG(u):f(v)=0,NG(v)∩V2={u}}. |
Clearly PN(u,f,G)≠∅ when f is a MRDF of G. A vertex u is called universal iff f(u)=2 for each MRDF f of G. A vertex u is called idle iff f(u)=0 for each MRDF f of G.
Proposition 2.1.[Rad et al. [6,11]] If u is universal or idle in graph G, then bR(G)≤dG(u). Moreover, γR(G)=γR(G−u) if u is idle.
Proposition 2.2. If u is idle in graph G, then bR(G)≤bR(G−u).
Proof. Let u be an idle vertex of graph G. Let B⊆E(G−u) be a minimum Roman bondage set of G−u. Then γR(G−u−B)>γR(G−u). We claim that B is also a Roman bondage set of G. Suppose to the contrary that γR(G−B)=γR(G). Then u is idle of G−B. By Proposition 2, we have γR(G−B)=γR(G−B−u). Hence γR(G−u−B)=γR(G)=γR(G−u), a contradiction with γR(G−u−B)>γR(G−u). Therefore bR(G)≤bR(G−u).
Proposition 2.3. Let T be a tree, N(u)=NT(u)={u1,u2,…,us} be the open neighborhood of u in T and Ti be the connected component of T−u that contains ui for each i=1,2,…,s. If u is idle in T and there exists a MRDF f of T such that at least min{s,3} vertices in N(u) can be assigned 2, then bR(T)=min{s,bR(Ti):i=1,2,…,s}.
Proof. Since u is idle in T, s≥1. By Proposition 2, γR(T)=γR(T−u)=∑si=1γR(Ti) and bR(T)≤min{dT(u),bR(T−u)}=min{s,bR(Ti):i=1,2,…,s}. If min{s,bR(Ti):i=1,2,…,s}=1, then bR(T)=1.
Next assume s≥2 and bR(Ti)≥2 for each i=1,2,…,s. Since there exists a MRDF f of T such that at least min{s,3}≥2 vertices in N(u) can be assigned 2, γR(T−uuj)=γR(T) for each positive integer j with 1≤j≤s. Let e∈E(Tj) for some 1≤j≤s, we have γR(Tj−e)=γR(Tj) since bR(Tj)≥2. Because there exists a MRDF f of T such that at least min{s,3}≥2 vertices in N(u) can be assigned 2, there exists some 1≤k≠j≤s and a MRDF f of T such that f(uk)=2. Then f|Ti is a MRDF of Ti for i≠j. Therefore γR(T−e)≤γR(Tj−e)+∑si=1,i≠jγR(Ti)=∑si=1γR(Ti)=γR(T). Hence bR(T)>1. If min{s,bR(Ti):i=1,2,…,s}=2, then bR(T)=2.
At last assume s≥3 and bR(Ti)=3 for each i=1,2,…,s. For any two different edges e1,e2∈E(T) and without loss of generality assume ej=uuij or ej∈E(Tij) for j=1,2 and 1≤ij≤s (admits i1=i2), we have γR(Tij−e1−e2)=γR(Tij) since bR(Tij)=3. Because there exists a MRDF f of T such that at least min{s,3}=3 vertices in N(u) can be assigned 2, there exists some 1≤k≠i1,i2≤s and a MRDF f of T such that f(uk)=2. Then f|Ti is a MRDF of Ti for i≠i1,i2. Therefore γR(T−e1−e2)≤γR(Ti1−e1)+γR(Ti2−e2)+∑si=1,i≠i1,i2γR(Ti)=∑si=1γR(Ti)=γR(T). Hence bR(T)≥2. Thus bR(T)=3.
We show a useful result in the following.
Theorem 3.1.[Rad et al. [6,11]] Let G be a graph and e=uv∈E(G). Then γR(G−e)>γR(G) iff f(u)=2 and v∈PN(u,f,G) or f(v)=2 and u∈PN(v,f,G) for each MRDF f of G.
Lemma 3.1. Let T be any tree. If there exists an universal vertex u∈V(G), then there exists v∈NT(u) such that γR(T−uv)>γR(T).
Proof. Let NT(u)={u1,u2,…,uk}. Suppose to the contrary that γR(T−uui)=γR(T) for each i=1,2,…,k. For each i=1,2,…,k, let Ti be the connected component of T−uui which contains ui and fi be a MRDF of T−uui. By definition, fi(V(Ti))=γR(Ti) for each i.
Let f be any MRDF of T. Then f(u)=2 since u is universal. If for all i∈{1,2,…,k}, fi(V(Ti))≤f(V(Ti)), then let
f′(x)={fi(x),x∈V(Ti),i=1,2,…,k;1,x=u. |
Then f′ is a RDF of T with f′(V(T))=∑kifi(V(Ti))+1<∑kif(V(Ti))+2=f(V(T)), a contradiction. Thus there exists some positive integer j with 1≤j≤k such that fj(V(Tj))>f(V(Tj)).
Let
f″(x)={f(x),x∈V(Tj);fj(x),otherwise. |
Since fj is a MRDF of T and u is universal, fj(u)=2 and hence f″ is also a RDF of T. However f″(V(T))=f(V(Tj))+fj(V−V(Tj))<fj(V(Tj))+fj(V−V(Tj))=fj(V(T))=γR(T), a contradiction. Therefore there exists v∈NT(u) such that γR(T−uv)>γR(T).
A vertex u is called free in G if any MRDF f have f(u)≠1 and there exist MRDFs f1 and f2 such that f1(u)=0 and f2(u)=2.
Proposition 2.2. Let e=uv be an edge in tree T. If both u and v are free vertices, and f(NT(u)∪NT(v)−{u,v})≤1 for any MRDF f. Then γR(T−uv)>γR(T).
Proof. Let f be any MRDF of T. Since both u and v are free vertices and f(NT(u)∪NT(v)−{u,v})≤1, {f(u),f(v)}={0,2} or f(u)=f(v)=2. We claim that {f(u),f(v)}={0,2}. Suppose to the contrary that there exists a MRDF f1 of T such that f1(u)=f1(v)=2. Then |PN(u,f1,T)|≥2 and |PN(v,f1,T)|≥2. Let NT(u)={v,u1,u2,…,uk}, k≥2 since |PN(u,f1,T)|≥2. There exists a MRDF f′1 of T such that f′1(u)=0 and f′1(v)=2 because u is a free vertex and f′1(N(u)∪N(v)−{u,v})≤1. Let Tu and Tv be the two connected components of T−uv that contain u and v, respectively. Note that f1(Tv)=f′1(Tv) and f1(Tu)=f′1(Tu). Let Tui be the connected component of T−uui that contains ui for each i=1,2,…,k. Since T is a tree, f′1(Tui)=γR(Tui)≥f1(T(ui)) for each i with 1≤i≤k.
![]() |
Since f1(Tu)=f′1(Tu), f1(u)=2 and f′1(u)=0, there exists some positive integer j with 1≤j≤k such that f′1(T(uj))=f1(T(uj))+2, or there exist two positive integers p and q with 1≤p,q≤k such that f′1(T(up))=f1(T(up))+1 and f′1(T(uq))=f1(T(uq))+1. If there exists some positive integer j with 1≤j≤k such that f′1(T(uj))=f1(T(uj))+2, then denote
f2(x)={f′1(x),x∈V(T−Tuj);f1(x),x∈V(Tuj)−uj;2,x=uj. |
Note that f2 is a Roman dominating function of T. Since f2(T)=f′1(T−Tuj)+f1(Tuj)+2=f′1(T−Tuj)+f′1(T(uj))=f′1(T), f2 is a MRDF of T. However, f2(uj)=2 is a contradiction with f2(NT(u)∪NT(v)−{u,v})≤1. If there exist two positive integers p and q with 1≤p,q≤k such that f′1(T(up))=f1(T(up))+1 and f′1(T(uq))=f1(T(uq))+1, then denote
f3(x)={f′1(x),x∈V(T−Tup−Tuq);f1(x),x∈V(Tup)∪V(Tuq)∖{up,uq};1,x∈{up,uq}. |
Note that f3 is a Roman dominating function of T. Since f3(T)=f′1(T−Tup−Tuq)+f1(Tup)+f1(Tuq)+1+1=f′1(T−Tup−Tuq)+f′1(T(up))+f′1(T(uq))=f′1(T), f3 is a MRDF of T. However, f3(up)=f3(uq)=1 is a contradiction with f3(NT(u)∪NT(v)−{u,v})≤1. Therefore {f(u),f(v)}={0,2} for any MRDF f of T.
Let f be any MRDF of T. Then {f(u),f(v)}={0,2}. Since f(NT(u)∪NT(v)−{u,v})≤1, v∈PN(u,f,T) if f(u)=2 or u∈PN(v,f,T) if f(v)=2. By Theorem 2, γR(T−uv)>γR(T).
Theorem 3.1. Let T be a tree. bR(T)=1 iff T has a universal vertex w, or
there exists an edge e=uv such that both u and v are free vertices, and f(NT(u)∪NT(v)−{u,v})≤1 for any MRDF f of T.
Proof. The sufficiency comes from Lemmas 3 and 3. Next we show the necessity. Assume there are no universal vertices in T. Since bR(T)=1, there exists an edge e=uv such that γR(T−uv)>γR(T). Let f be any MRDF of T. By Theorem 2, f(u)=2 and v∈PN(u,f,T) or f(v)=2 and u∈PN(v,f,T). We have both u and v are free vertices since both of them are not universal vertices.
We only need to show f(NT(u)∪NT(v)−{u,v})≤1. Suppose to the contrary that f(NT(u)∪NT(v)−{u,v})≥2. Without loss of generality assume f(u)=2 and v∈PN(u,f,T). Since u and v are free vertices, there exists a MRDF f′ of T such that f′(v)=2 and u∈PN(v,f′,T). We claim that f(NT(u))=0. Otherwise there exists a vertex w∈N(u) such that f(w)=2 since f(u)=2. Let Tw be the connected component of T−uw which contains w. Then f|Tw is a MRDF of Tw. Also f′|Tw is a MRDF of Tw. Denote
f″(x)={f(x),x∈V(Tw);f′(x),otherwise. |
Clearly f″ is a MRDF of T. However, f″(u)=0 and f″(v)=f″(w)=2 is a contradiction with u∈PN(v,f″,T) by Theorem 2. Thus f(NT(u))=0 and f(NT(v)−u)≥2. Since v∈PN(u,f,T), there exists at least two vertices s and t in NT(v)−u such that f(s)=f(t)=1. Denote
f1(x)={f(x),x∈V(T)∖{v,s,t};2,x=v;0,x∈{s,t}. |
Clearly f1 is a MRDF of T. However, f1(u)=f1(v)=2, a contradiction with v∈PN(u,f,T). Therefore f(NT(u)∪NT(v)−{u,v})≤1.
We characterize trees with Roman bondage number 1 in the above paragraph. Since bR(T)≤3 for tree T, we have tried to obtain the similar results for bR(T) equals to 2 or 3. Unfortunately, it seems very difficult or we can not get similar results for bR(T) equals to 2 or 3. Indeed, it may be much easier to deal with bR(T)=3. But the similar method does not work. We will try to find other method to study the cases of bR(T) equals to 2 or 3.
The authors would like to express their gratitude to the anonymous referees for their critical commons and helpful suggestions on the original manuscript, which resulted in this version.
This research was supported by NNSF of China (No. 11401004) and Anhui Provincial Natural Science Foundation (No. 1708085MA01).
The authors declare no conflict of interest.
[1] |
S. Akbari, M. Khatirinejadand, S. Qajar, A note on Roman bondage number of planar graphs, Graph. Combinator., 29 (2013), 327-331. doi: 10.1007/s00373-011-1129-8
![]() |
[2] | A. Bahremandpour, F. T. Hu, S. M. Sheikholeslami, et al. Roman bondage number of a graph, Discrete Math. Algorithm. Appl., 5 (2013), 1-15. |
[3] | X. G. Chen, A note on the double Roman domination number of graphs, Czech. Math. J., 70 (2020), 205-212. |
[4] |
E. J. Cockayne, P. A. Dreyer, S. M. Hedetniemi, et al. Roman domination in graphs, Discrete Math., 278 (2004), 11-22. doi: 10.1016/j.disc.2003.06.004
![]() |
[5] |
J. F. Fink, M. S. Jacobson, L. F. Kinch, et al. The bondage number of a graph, Discrete Math., 86 (1990), 47-57. doi: 10.1016/0012-365X(90)90348-L
![]() |
[6] | A. Hansberg, N. J. Rad, L. Volkmann, Vertex and edge critical Roman domination in graphs, Utliltas Math., 92 (2013), 73-88. |
[7] | T. W. Haynes, S. T. Hedetniemi, P. J. Slater, Fundamentals of domination in graphs, New York: Marcel Dekker, 1998. |
[8] | T. W. Haynes, S. T. Hedetniemi, P. J. Slater, Domination in graphs: Advanced topics, New York: Marcel Dekker, 1998. |
[9] |
N. J. Rad, L. Volkmann, Roman bondage in graphs, Discuss. Math. Graph T., 31 (2011), 763-773. doi: 10.7151/dmgt.1578
![]() |
[10] |
N. J. Rad, L. Volkmann, On the Roman bondage number of planar graphs, Graph. Combinator., 27 (2011), 531-538. doi: 10.1007/s00373-010-0978-x
![]() |
[11] | N. J. Rad, L. Volkmann, Changing and unchanging the Roman domination number of a graph, Utliltas Math., 89 (2012), 79-95. |
[12] | V. Samodivkin, On the Roman bondage number of graphs on surfaces, Int. J. Graph Theory Appl., 1 (2015), 67-75. |
[13] |
I. Stewart, Defend the Roman empire, Sci. Am., 281 (1999), 136-138. doi: 10.1038/scientificamerican1299-136
![]() |
[14] | J. M. Xu, Theory and application of graphs, Dordrecht/Boston/London: Kluwer Academic Publishers, 2003. |
[15] | J. M. Xu, On bondage numbers of graphs: A survey with some comments, Int. J. Combinator., 2013 (2013), 1-34. |