
Let G=(V,E) be a simple graph with vertex set V and edge set E, and let f be a function f:V↦{0,1,2}. A vertex u with f(u)=0 is said to be undefended with respect to f if it is not adjacent to a vertex with positive weight. The function f is a weak Roman dominating function (WRDF) if each vertex u with f(u)=0 is adjacent to a vertex v with f(v)>0 such that the function fu:V↦{0,1,2}, defined by fu(u)=1, fu(v)=f(v)−1 and fu(w)=f(w) if w∈V−{u,v}, has no undefended vertex. The weight of f is w(f)=∑v∈Vf(v). The weak Roman domination number, denoted γr(G), is the minimum weight of a WRDF in G. The domination number, denoted γ(G), is the minimum cardinality of a dominating set in G. In this paper, we give some sufficient conditions for a tree to have its weak Roman domination number be equal to its domination number plus 1 (γr(T)=γ(T)+1) by recursion and construction.
Citation: 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[J]. AIMS Mathematics, 2023, 8(8): 17702-17718. doi: 10.3934/math.2023904
[1] | Fu-Tao Hu, Xing Wei Wang, Ning Li . Characterization of trees with Roman bondage number 1. AIMS Mathematics, 2020, 5(6): 6183-6188. doi: 10.3934/math.2020397 |
[2] | 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 |
[3] | 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 |
[4] | 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 |
[5] | Zepeng Li . A note on the bounds of Roman domination numbers. AIMS Mathematics, 2021, 6(4): 3940-3946. doi: 10.3934/math.2021234 |
[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] | 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 |
[8] | 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 |
[9] | Mingyu Zhang, Junxia Zhang . On Roman balanced domination of graphs. AIMS Mathematics, 2024, 9(12): 36001-36011. doi: 10.3934/math.20241707 |
[10] | Abel Cabrera-Martínez, Andrea Conchado Peiró . On the {2}-domination number of graphs. AIMS Mathematics, 2022, 7(6): 10731-10743. doi: 10.3934/math.2022599 |
Let G=(V,E) be a simple graph with vertex set V and edge set E, and let f be a function f:V↦{0,1,2}. A vertex u with f(u)=0 is said to be undefended with respect to f if it is not adjacent to a vertex with positive weight. The function f is a weak Roman dominating function (WRDF) if each vertex u with f(u)=0 is adjacent to a vertex v with f(v)>0 such that the function fu:V↦{0,1,2}, defined by fu(u)=1, fu(v)=f(v)−1 and fu(w)=f(w) if w∈V−{u,v}, has no undefended vertex. The weight of f is w(f)=∑v∈Vf(v). The weak Roman domination number, denoted γr(G), is the minimum weight of a WRDF in G. The domination number, denoted γ(G), is the minimum cardinality of a dominating set in G. In this paper, we give some sufficient conditions for a tree to have its weak Roman domination number be equal to its domination number plus 1 (γr(T)=γ(T)+1) by recursion and construction.
For a graph G=(V,E), let f:V↦{0,1,2}, and let (V0, V1, V2) be the ordered partition of V induced by f, where Vi={v∈V | f(v)=i} and |Vi|=ni, for i=0,1,2. Note that there exists a 1−1 correspondence between the functions f:V↦{0,1,2} and the ordered partitions (V0, V1, V2) of V. Thus, we will write f=(V0,V1,V2).
Motivated by an article in Scientific American by I. Stewart [1] entitled "Defend the Roman Empire!", E. J. Cockayne et al. [2] defined a Roman dominating function (RDF) on a graph G=(V,E) to be a function f:V↦{0,1,2} satisfying the condition that every vertex u for which f(u)=0 is adjacent to at least one vertex v for which f(v)=2. For a real-valued function f:V↦R, the weight of f is w(f)=∑v∈Vf(v)=|V1|+2|V2|, and for S⊆V we define f(S)=∑v∈Sf(v), so w(f)=f(V). The Roman domination number, denoted γR(G), is the minimum weight of a RDF in G; that is, γR(G) = min{w(f) | f is a RDF in G}. For example, the Roman domination number of the 2×5 grid graph G2,5 is 6 (Figure 1(a)). A RDF of weight γR(G) we call a γR(G)-function. Roman domination in graphs has been studied, for example, in [2,3].
This definition of a Roman dominating function is motivated as follows. Each vertex in our graph represents a location in the Roman Empire. A location (vertex v) is considered unsecured if no legions are stationed there (i.e., f(v)=0) and secured otherwise (i.e., if f(v)∈{1,2}). An unsecured location (vertex v) can be secured by sending a legion to v from an adjacent location (an adjacent vertex u). However, Emperor Constantine the Great, in the fourth century A.D., decreed that a legion cannot be sent from a secured location to an unsecured location if doing so leaves that location unsecured (i.e., if f(v)=1). Thus, two legions must be stationed at a location (f(v)=2) before one of the legions can be sent to an adjacent location. In this way, Emperor Constantine the Great can defend the Roman Empire. Since it is expensive to maintain a legion at a location, the Emperor would like to station as few legions as possible, while still defending the Roman Empire. A Roman dominating function of weight γR(G) corresponds to such an optimal assignment of legions to locations.
M. A. Henning and S. T. Hedetniemi [4] explored a new strategy of defending the Roman Empire that has the potential of saving the Emperor Constantine the Great substantial costs of maintaining legions, while still defending the Roman Empire (from a single attack). Let G=(V,E) be a graph and f be a function f=(V0,V1,V2). A vertex u∈V0 is undefended with respect to f, or simply undefended if the function f is clear from the context, if it is not adjacent to a vertex in V1 or V2. The function f is a weak Roman dominating function (WRDF) if each vertex u∈V0 is adjacent to a vertex v∈V1∪V2 such that the function fu:V↦{0,1,2}, defined by fu(u)=1, fu(v)=f(v)−1 and fu(w)=f(w) if w∈V−{u,v}, has no undefended vertex. The weak Roman domination number, denoted by γr(G), is the minimum weight of a WRDF in G, that is, γr(G) = min{w(f) | f is a WRDF in G}. For example, the weak Roman domination number of the 2×5 grid graph G2,5 is 4 (Figure 1(b)). A WRDF of weight γr(G) we call a γr(G)-function. Weak Roman domination in graphs has been studied, for example, in [4,5,6,7].
This definition of a WRDF is motivated as follows. Using notation introduced earlier, we define a location to be undefended if the location and every location adjacent to it are unsecured (i.e., have no legion stationed there). Since an undefended location is vulnerable to an attack, we require that every unsecure location be adjacent to a secure location in such a way that the movement of a legion from the secure location to the unsecure location does not create an undefended location. Hence, every unsecure location can be defended without creating an undefended location. In this way Emperor Constantine the Great can still defend the Roman Empire. Such a placement of legions corresponds to a WRDF, and a minimum such placement of legions corresponds to a minimum WRDF.
Roman domination is a typical control problem with rich historical background and mathematical background [2,4,8,9,10,11,12,13,14,15]. E. J. Cockayne et al. [2] studied the Roman domination of graphs, including determining the Roman domination numbers of the paths, cycles, complete n-partite graphs and 2×n grid graphs; giving some upper and lower bounds of the Roman domination number; and characterizing the graphs with γR(G)=γ(G), γR(G)=γ(G)+1, γR(G)=γ(G)+2 and the trees with γR(T)=γ(T)+1, γR(T)=γ(T)+2, etc. M. A. Henning and S. T. Hedetniemi [4] studied the weak Roman domination of graphs, including determining the weak Roman domination numbers of paths and cycles and characterizing the graphs with γr(G)=γ(G), etc. Authors of this paper have also studied the weak Roman domination of graphs, including characterizing the trees with γr(T)=γ(T) in [6] and determining that the path P3, stars K1,t(t≥2) and trees T which consist of the center vertices of stars K1,t1, K1,t2, ⋯, K1,tn(ti≥3,i=1,2,⋯,n) to form a path are weak Roman graphs in [7]. In this paper, we give some sufficient conditions for a tree to have its weak Roman domination number be equal to its domination number plus 1 (γr(T)=γ(T)+1) by recursion and construction. The graphs considered in this paper are finite non-trivial simple graphs.
Let G=(V,E) be a graph with vertex set V of order n and edge set E, and let v be a vertex in V. The degree of the vertex v is denoted as d(v). A graph is called complete if every pair of different vertices is connected by exactly one edge, denoted by Kn. (where, K1 represents a complete graph with vertex set V of order 1 and K2 represents a complete graph with vertex set V of order 2.) A nonempty sequence of alternating vertices and edges W=v0e1v1e2v2⋯ekvk (where v0,v1,⋯,vk are different) in a graph G is called a path, denoted by Pn. The open neighborhood of v is N(v)={u∈V|uv∈E}, and the closed neighborhood of v is N[v]={v}∪N(v). For a set S⊆V, its open neighborhood N(S)=∪v∈SN(v)−S, and its closed neighborhood N[S]=N(S)∪S.
Let G=(V,E) be a graph, and let S⊆V. A set S dominates a set U, denoted S≻U, if every vertex in U is adjacent to a vertex of S. If S≻V−S, then S is called a dominating set of G [16]. The domination number γ(G) is the minimum cardinality of a dominating set of G, namely, γ(G)=min{|S||S≻V−S}. A dominating set of cardinality γ(G) we call a γ(G)-set.
Let T=(V,E) be a tree with vertex set V of order n and edge set E. A leaf of T is a vertex of degree 1, while a support vertex of T is a vertex adjacent to a leaf. A strong support vertex is adjacent to at least two leaves. In this paper, we denote the set of all support vertices of T by S(T), the set of all strong support vertices by SS(T) and the set of leaves by L(T). If S⊆V, and for all vertices in S, as long as there is an edge in the tree T, this edge appears in the subtree T[S], then T[S] is said to be the subtree of tree T induced by S, denoted as T[S]. For a positive integer t, the complete bipartite graph K1,t is called a star, the vertex whose degree is t (t≥2) is the center vertex, and a vertex whose degree is 1 is an outer vertex.
Previously known results on domination number and weak Roman domination number are the following.
Lemma1. [4] For any graph G, γ(G)≤γr(G)≤γR(G)≤2γ(G).
Lemma2. [4] For any n≥1, γr(Pn)=⌈3n7⌉. For any n≥4, γr(Cn)=⌈3n7⌉.
Lemma3. [6] For any tree T with γr(T)=γ(T), tree T does not contain any strong support vertex, that is, ∀v∈V,v∉SS(T).
Lemma4. [17] For any n≥1, γ(Pn)=⌈n3⌉.
Lemma5. [6] Let T=(V,E) be a tree with vertex set V of order n and edge set E. Then, γr(T)=γ(T) if and only if one of the following is true:
Case 1. m (m is a positive integer) is even, the tree T contains m2 K2 (let ui,vi∈V, uivi∈E,i=1,2,⋯,m2), and the following conditions are satisfied:
(a) Each K2 has a vertex ui∈S(T) and another vertex vi∈L(T), S(T)={u1,u2,⋯,um2}, and L(T)={v1,v2,⋯,vm2}.
(b) ∀u∈S(T), except for its leaf, u is adjacent to other vertices in S(T).
Case 2. The tree T does not contain any strong support vertex and contains m (m is a positive integer) K2 (let ui,vi∈V, uivi∈E,i=1,2,⋯,m) and n (n is a positive integer) stars K1,t1,K1,t2,⋯,K1,tn (ti≥2,i=1,2,⋯,n) (the set formed by their center vertices is C, and the set formed by their outer vertices is W), and the following conditions are satisfied:
(a) Each K2 has a vertex ui∈S(T) and another vertex vi∈L(T), S(T)={u1,u2,⋯,um}, and L(T)={v1,v2,⋯,vm}.
(b) ∀w∈W, there is a unique support vertex u∈S(T), such that wu∈E(T), and ∀v∈C, v is at least 2 away from every vertex in S(T).
(c) ∀u∈S(T), except for its leaf, either u is adjacent to an outer vertex w∈W, or u is adjacent to other vertices in S(T).
Theorem1. For any tree T, let the set S be any γ(T)-set of the tree T. Then, SS(T)⊆S.
Proof. Suppose the contrary, that is, there is v∈SS(T)∖S, and let u1,u2,⋯, ul (l≥2) be the leaves of v. Because v∉S, u1,u2,⋯,ul∈S (l≥2). Let S1=S∪{v}∖{u1,u2,⋯,ul (l≥2)}. Then, obviously, S1 is a dominating set of the tree T, and |S1|=|S∪{v}∖{u1,u2,⋯,ul (l≥2)}|≤|S|−1. This contradicts S being the γ(T)-set of the tree T. Therefore, SS(T)⊆S.
Theorem2. For any tree T, there is a γ(T)-set S such that S(T)⊆S.
Proof. From Theorem 1 we know SS(T)⊆S. Let S0 be a γ(T)-set of the tree T=(V,E). Then, SS(T)⊆S0. If S(T)∖SS(T)⊆S0, then the conclusion is true.
Otherwise, let v∈S(T)∖S0, and v∉SS(T). Then, let u be the only leaf of v, that is, uv∈E, and d(u)=1. Thus, N(u)={v}, but v∉S0, so u∈S0. Let S1=S0∪{v}∖{u}. Then, obviously, S1 is a dominating set of the tree T. Also, |S1|=|S0|=γ(T), so S1 is also a γ(T)-set of the tree T, and the number of support vertices in S1 is 1 more than that in S0.
If S(T)∖SS(T)⊆S1, then the conclusion is true. Otherwise, ∃v1∈S(T)∖S1, and v1∉SS(T). Similar to the above, find S2 such that S2 is also a γ(T)-set of the tree T, and the number of support vertices in S2 is 1 more than that in S1.
⋯⋯ |
Repeat the above operation. Since the tree T is finite, the number of support vertices must also be finite. Each repetition will increase the number of support vertices by 1, so a γ(T)-set Sk must exist after a finite number of repetitions such that S(T)∖SS(T)⊆Sk. Hence, the conclusion follows.
Theorem3. For any tree T=(V,E), there is a γr(T)-function f=(V0,V1,V2), such that SS(T)⊆V2.
Proof. Suppose there is v∈SS(T)∖V2, and then v∈V0∪V1. Because v∈SS(T), v has at least two leaves, denoted u1,u2,⋯,ul (l≥2). If v∈V0, since u1,u2,⋯,ul (l≥2) are only adjacent to v, u1,u2,⋯, ul∈V1. Let fv=(V∗0,V∗1,V∗2)=(V0∪{u1,u2,⋯,ul}∖{v},V1∖{u1,u2,⋯,ul},V2∪{v}). Then, fv has no undefended vertex, and w(fv)=|V1∖{u1,u2,⋯,ul}|+2|V2∪{v}|=|V1|−l+2|V2|+2=|V1|+2|V2|+2−l≤|V1|+2|V2|=w(f). So, fv is also a γr(T)-function, and v∈V∗2.
If v∈V1, at most one of u1,u2,⋯,ul (l≥2) belongs to V0 (otherwise, if u1,u2∈V0, send a legion from v to u1 for security defense, such that u2 is an undefended vertex, a contradiction), denoted by u1∈V0. Then, u2,u3,⋯,ul∈V1. Let fv=(V∗0,V∗1,V∗2)=(V0∪{u2,u3,⋯,ul},V1∖{v,u2,u3,⋯,ul},V2∪{v}). Then, fv has no undefended vertex, and w(fv)=|V1∖{v,u2,u3,⋯,ul}|+2|V2∪{v}|=|V1|−l+2|V2|+2=|V1|+2|V2|+2−l≤|V1|+2|V2|=w(f). So, fv is also a γr(T)-function, and v∈V∗2. Hence, there is a γr(T)-function f=(V0,V1,V2), such that SS(T)⊆V2.
Theorem4. For any tree T with γr(T)=γ(T)+1, tree T contains at most 1 strong support vertex.
Proof. Let T=(V,E) be a tree with γr(T)=γ(T)+1, and let f=(V0,V1,V2) be a γr(T)-function of the tree T. V1∪V2≻V0, so γ(T)≤|V1|+|V2|, and then |V2|≤1 (otherwise, if |V2|≥2, then γr(T)=|V1|+2|V2|=|V1|+|V2|+|V2|≥|V1|+|V2|+2≥γ(T)+2>γ(T)+1, which contradicts the assumptions).
Suppose the contrary, that is, T has at least two strong support vertices, say, v1,v2∈SS(T). By Theorem 3, there is a γr(T)-function f=(V0,V1,V2) such that SS(T)⊆V2, and then v1,v2∈V2. Therefore, |V2|≥2, a contradiction. So, tree T contains at most 1 strong support vertex.
Theorem5. If the tree T=(V,E) has one and only one strong support vertex, and there is a γ(T)-set S such that every vertex in S is a support vertex, and G[S] is a tree, as shown in Figure 2, then, γr(T)=γ(T)+1.
Proof. By Theorem 2, there is a γ(T)-set S1 such that S(T)⊆S1. There is also a γ(T)-set S such that every vertex in S is a support vertex, and then S⊆S(T)⊆S1. Since, |S|=|S1|, S=S(T)=S1, and γ(T)=|S(T)|.
The tree T=(V,E) has one and only one strong support vertex, say, v∗∈SS(T), and set l1,l2,⋯,lt (t≥2) for its leaves. Let f=(V0,V1,V2)=(V−S(T),S(T)∖{v∗},{v∗}). Then, it has no undefended vertex, as shown in Figure 2. Since G[S(T)] is a tree, ∀u,v∈S(T), there is no u−v path that contains a vertex in V−S(T) as the inner vertex. (Otherwise, if there is such a road P1, since G[S(T)] is the tree, the u−v path P2 exists in G[S(T)]. Obviously, P1≠P2, so T must contain cycles, which is a contradiction.) Since S(T) is a dominating set, the tree T has one and only one strong support vertex. Then, ∀ u∈V−S(T)∪{l1,l2,⋯,lt}, u can only be adjacent to the only vertex v in S(T)∖{v∗}, and d(u)=1. (Otherwise, if d(u)≥2, let wu∈E(T). However, w∉S(T), and w∉N[S(T)]. This contradicts S(T) being a dominating set.) So, ∀ u∈V−S(T)∪{l1,l2,⋯,lt}, let fu=(V∗0,V∗1,V∗2)=(V−S(T)∪{u}∖{v},S(T)∪{u}∖{v,v∗},{v∗}). Obviously, no undefended vertices will be created. l1,l2,⋯,lt (t≥2) can be v∗ security defense, so f is a WRDF of the tree T. Therefore, γr(T)≤w(f)=|S(T)∖{v∗}|+2|{v∗}|=|S(T)|−1+2=|S(T)|+1=γ(T)+1.
From Lemma 3 and the fact that T has a strong support vertex, γ(T)≠γr(T). Again, by Lemma 1, know γ(T)≤γr(T), and then γr(T)>γ(T), or γr(T)≥γ(T)+1. Hence, γr(T)=γ(T)+1.
Theorem6. Suppose the tree T=(V,E) has only one strong support vertex (say, v∗∈SS(T), and set l1,l2,⋯,ls (s≥2) for its leaves) and contains m (m is a positive integer) K2 (let ui,vi∈V, uivi∈E,i=1,2,⋯,m) and n (n is a positive integer) stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n) (the set formed by their center vertices is C, and the set formed by their outer vertices is W) and the following conditions are satisfied:
(a) Each K2 has a vertex ui∈S(T) and another vertex vi∈L(T), S(T)={u1,u2,⋯,um,v∗}, and L(T)={v1,v2,⋯,vm}∪{l1,l2,⋯,ls}.
(b) ∀w∈W, there is a unique support vertex u∈S(T), such that wu∈E(T), and ∀v∈C, v is at least 2 away from every vertex in S(T).
(c) ∀u∈S(T), except for its leaf, either u is adjacent to an outer vertex w∈W, or u is adjacent to other vertices in S(T).
As shown in Figure 3, then, γr(T)=γ(T)+1.
Proof. By Theorem 2, there is a γ(T)-set S such that S(T)⊆S. Based on assumptions, the set formed by the center vertices of stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n) is C. Then, obviously, S(T)∪C is a dominating set of the tree T, and S(T)∩C=∅, so γ(T)≤|S(T)∪C|=|S(T)|+|C|. Based on assumptions, ∀v∈C, v is at least 2 away from every vertex in S(T), so γ(T)≥|S(T)|+|C|. Hence, γ(T)=|S(T)|+|C|, and thus S(T)∪C is a γ(T)-set of the tree T.
The tree T=(V,E) has one and only one strong support vertex, say, v∗∈SS(T), and set l1,l2,⋯,ls (s≥2) for its leaves. Let f=(V0,V1,V2)=(V−S(T)∪C,S(T)∪C∖{v∗},{v∗}), and then it has no undefended vertex, as shown in Figure 3. It is clear that the vertices in V−S(T)∪C consist of the set of all leaves L(T) and the set of outer vertices W of the stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n), that is, V−S(T)∪C=L(T)∪W, and L(T)∩W=∅. ∀ v∈V−S(T)∪C=L(T)∪W, if v∈L(T)∖{l1,l2,⋯,ls}, let uv∈E, and u∈S(T). Then, obviously, fv=(V∗0,V∗1,V∗2)=(L(T)∪W∪{u}∖{v},S(T)∪C∪{v}∖{u,v∗},{v∗}) has no undefended vertex. If v∈{l1,l2,⋯,ls}, since f(v∗)=2, then send a legion from v∗ to v for security defense without creating undefended vertices. If v∈W, let uv∈E, and u∈C. Then, obviously, fv=(V∗0,V∗1,V∗2)=(L(T)∪W∪{u}∖{v},S(T)∪C∪{v}∖{u,v∗},{v∗}) has no undefended vertex. So, f is a WRDF of the tree T. Therefore, γr(T)≤w(f)=|S(T)∪C∖{v∗}|+2|{v∗}|=|S(T)∪C|−1+2=|S(T)|+|C|+1=γ(T)+1.
From Lemma 3 and the fact that T has a strong support vertex, γ(T)≠γr(T). Again, by Lemma 1, know γ(T)≤γr(T), and then γr(T)>γ(T), or γr(T)≥γ(T)+1. Hence, γr(T)=γ(T)+1.
Theorem7. Suppose the tree T=(V,E) contains a star K∗1,t (t≥3) (its center vertex is denoted as v∗; its outer vertex is denoted as w1,w2,⋯,wt (t≥3), among them, there is at least one vertex that is a leaf of the tree T, denoted w1,w2,⋯,ws; and at least two outer vertices are adjacent to the support vertex of the tree T, denoted ws+1,ws+2,⋯,wt (t≥3)) and contains m (m is a positive integer) K2 (let ui,vi∈V, uivi∈E,i=1,2,⋯,m) and n (n is a positive integer) stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n) (the set formed by their center vertices is C, and the set formed by their outer vertices is W), the subtree T−K∗1,t does not contain any strong support vertices, and the following conditions are satisfied:
(a) Each K2 has a vertex ui∈S(T) and another vertex vi∈L(T), S(T)={u1,u2,⋯,um,v∗}, and L(T)={v1,v2,⋯,vm}∪{w1,w2,⋯,ws}.
(b) ∀w∈W∪{ws+1,ws+2,⋯,wt}, there is a unique support vertex u∈S(T), such that wu∈E(T), and ∀v∈C, v is at least 2 away from every vertex in S(T).
(c) ∀u∈S(T)∖{v∗}, except for its leaf, either u is adjacent to an outer vertex w∈W∪{ws+1,ws+2,⋯,wt}, or u is adjacent to other vertices in S(T)∖{v∗}.
As shown in Figure 4, then, γr(T)=γ(T)+1.
Proof. By Theorem 2, there is a γ(T)-set S such that S(T)⊆S. Based on assumptions, the set formed by the center vertices of the stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n) is C. Then, obviously, S(T)∪C is a dominating set of the tree T, and S(T)∩C=∅, so γ(T)≤|S(T)∪C|=|S(T)|+|C|. Based on assumptions, ∀v∈C, v is at least 2 away from every vertex in S(T), so γ(T)≥|S(T)|+|C|. Hence, γ(T)=|S(T)|+|C|, and thus S(T)∪C is a γ(T)-set of the tree T.
Let f=(V0,V1,V2)=(V−S(T)∪C,S(T)∪C∖{v∗},{v∗}). Then, it has no undefended vertex, as shown in Figure 4. Obviously, the vertices in V−S(T)∪C consist of three parts: (i) the set of all leaves L(T) (where w1,w2,⋯,ws∈L(T)); (ii) the set of outer vertices W of the stars K1,t1,K1,t2,⋯,K1,tn (ti≥2,i=1,2,⋯,n); (iii) the set of outer vertices {ws+1,ws+2,⋯,wt} of the star K∗1,t (t≥3), that is, V−S(T)∪C=L(T)∪W∪{ws+1,ws+2,⋯,wt}, and L(T)∩W=∅, L(T)∩{ws+1,ws+2,⋯,wt}=∅, W∩{ws+1,ws+2, ⋯,wt}=∅. ∀ v∈V−S(T)∪C=L(T)∪W∪{ws+1,ws+2,⋯,wt}, if v∈L(T)∖{w1,w2,⋯,ws}, let uv∈E, and u∈S(T). Then, obviously, fv=(V∗0,V∗1,V∗2)=(L(T)∪W∪{ws+1,ws+2,⋯, wt}∪{u}∖{v}, S(T)∪C∪{v}∖{u,v∗},{v∗}) has no undefended vertex. If v∈{w1,w2,⋯,wt}, since f(v∗)=2, then send a legion from v∗ to v for security defense without creating undefended vertices. If v∈W, let uv∈E, and u∈C. Then, obviously, fv=(V∗0,V∗1,V∗2)=(L(T)∪W∪{ws+1,ws+2,⋯,wt}∪{u}∖{v}, S(T)∪C∪{v}∖{u,v∗},{v∗}) has no undefended vertex. So, f is a WRDF of the tree T. Therefore, γr(T)≤w(f)=|S(T)∪C∖{v∗}|+2|{v∗}|=|S(T)∪C|−1+2=|S(T)|+|C|+1=γ(T)+1.
Since at least one outer vertex of the star K∗1,t (t≥3) is a leaf of the tree T, then for any weak Roman dominating function f=(V0,V1,V2) of the tree T, f(N[v∗])≥2. (Otherwise, if f(N[v∗])≤1, then f(v∗)=1, and f(w1)=0. Also, since the outer vertex ws+1 is adjacent to the support vertex of the tree T, send a legion from v∗ to ws+1 for security defense, such that w1 is an undefended vertex, which is a contradiction.) On the other hand, {v∗}≻N(v∗), so γr(T)≥γ(T)+1. Hence, γ(T)=γr(T)+1.
Theorem8. Suppose the tree T=(V,E) does not contain any strong support vertices, contains m (m is a positive integer) K2 (let ui,li∈V, uili∈E,i=1,2,⋯,m) and has and only has one of the following seven cases: (1) K1; (2) P2; (3) P4; (4) P5; (5) P6; (6) P7; (7) P9. Further, suppose the following conditions are satisfied:
(a) Each K2 has a vertex ui∈S(T) and another vertex li∈L(T), S(T)={u1,u2,⋯,um}, and L(T)={l1,l2,⋯,lm}.
(b) Both ends of the seven cases: (1) K1; (2) P2; (3) P4; (4) P5; (5) P6; (6) P7; (7) P9 are adjacent to a unique support vertex u∈S(T).
(c) ∀u∈S(T), except for its leaf, either u is adjacent to one of the endpoints of the seven cases: (1) K1; (2) P2; (3) P4; (4) P5; (5) P6; (6) P7; (7) P9, or u is adjacent to other vertices in S(T).
As shown in Figure 5, then, γr(T)=γ(T)+1.
Proof. By Theorem 2, there is a γ(T)-set S such that S(T)⊆S. Then, divided into seven cases, the proof is as follows:
(1) If the tree T has and only has the K1, as shown in Figure 5(T1), based on assumptions, obviously, S(T1) can dominate the tree T1, then S⊆S(T1). Also, S(T1)⊆S, so S=S(T1), and γ(T1)=|S(T1)|.
Let f=(V0,V1,V2)=(V−S(T1)∪{v1},S(T1)∪{v1},∅). Then, it has no undefended vertex. Since the tree T1 does not contain any strong support vertices, both ends of K1 are adjacent to a unique support vertex, and S(T1) is a dominating set of T1, ∀u∈V−S(T1)∪{v1}=L(T1), u can only be adjacent to the only vertex v in S(T1), and d(u)=1. Let fu=(V∗0,V∗1,V∗2)=(V−S(T1)∪{v1,u}∖{v},S(T1)∪{v1,u}∖{v},∅). Then, obviously, fu has no undefended vertex, so f is a WRDF of the tree T1. Therefore, γr(T1)≤w(f)=|S(T1)∪{v1}|=|S(T1)|+1=γ(T1)+1.
Since the leaves of T1 one-to-one correspond to their support vertices, and each leaf requires at least one legion for security defense or security defense by its support vertex, and if the vertex v1 for security defense, either f(v1)=1, or f(u1)=2, or f(u2)=2, as shown in Figure 5(T1). So, γr(T1)≥|S(T1)|+1=γ(T1)+1. Therefore, γr(T1)=γ(T1)+1.
(2) If the tree T has and only has the path P2, as shown in Figure 5(T2), the proof of (1) is the same.
(3) If the tree T has and only has the path P6, as shown in Figure 5(T5), based on assumptions, S(T5)∪{v2,v5} can dominate the tree T5, so γ(T5)≤|S(T5)∪{v2,v5}|=|S(T5)|+2. v1 is adjacent to u1, and v6 is adjacent to u2. u1,u2∈S(T5), and from Lemma 4, γ(P4)=⌈43⌉=2. So, γ(T5)≥|S(T5)|+2. Therefore, γ(T5)=|S(T5)|+2.
Let f=(V0,V1,V2)=(V−S(T5)∪{v2,v4,v5},S(T5)∪{v2,v4,v5},∅). Then, it has no undefended vertex. The tree T5 does not contain any strong support vertices, both ends of P6 are adjacent to a unique support vertex, and S(T5)∪{v2,v5} is a dominating set of T5. Then, ∀u∈V−S(T5)∪{v2,v4,v5}=L(T5)∪{v1,v3,v6}, if u∈L(T5), u can only be adjacent to the only vertex v in S(T5) and d(u)=1. Let fu=(V∗0,V∗1,V∗2)=(V−S(T5)∪{v2,v4,v5,u}∖{v},S(T5)∪{v2,v4,v5,u}∖{v},∅). Then, obviously, fu has no undefended vertex. If u=v1, let fu=(V∗0,V∗1,V∗2)=(V−S(T5)∪{v1,v4,v5},S(T5)∪{v1,v4,v5},∅), and then, obviously, fu has no undefended vertex. If u=v3, let fu=(V∗0,V∗1,V∗2)=(V−S(T5)∪{v2,v3,v5},S(T5)∪{v2,v3,v5},∅). Then, obviously, fu has no undefended vertex. If u=v6, let fu=(V∗0,V∗1,V∗2)=(V−S(T5)∪{v2,v4,v6},S(T5)∪{v2,v4,v6},∅), and then, obviously, fu has no undefended vertex. So, f is a WRDF of the tree T5. Therefore, γr(T5)≤w(f)=|S(T5)∪{v2,v4,v5}|=|S(T5)|+3=γ(T5)+1.
Since the leaves of T5 one-to-one correspond to their support vertices, and each leaf requires at least one legion for security defense or security defense by its support vertex, and if the vertex v1 for security defense, either f(v1)=1, or f(v2)=1, or f(u1)=2. The same if the vertex v6 for security defense, either f(v6)=1, or f(v5)=1, or f(u2)=2, as shown in Figure 5(T5). That is, the vertex v1 cannot be securely defended by u1, and the vertex v6 cannot be securely defended by u2, unless u1 or u2 are stationed in two legions. From Lemma 2, γr(P6)=⌈3×67⌉=3, so γr(T5)≥|S(T5)|+3=γ(T5)+1. Therefore, γr(T5)=γ(T5)+1.
(4) If the tree T has and only has the path P4, P5, P7, P9, as shown in Figure 5(T3, T4, T6, T7), the proof of (3) is the same.
Note1. Suppose the tree T=(V,E) does not contain any strong support vertices, contains m (m is a positive integer) K2 (let ui,li∈V, uili∈E,i=1,2,⋯,m) and has only one of the following three cases: (1) P3; (2) P8; (3) Pn (n≥10). Suppose also that the following conditions are satisfied: (a) Each K2 has a vertex ui∈S(T) and another vertex li∈L(T), S(T)={u1,u2,⋯,um}, and L(T)={l1,l2,⋯,lm}. (b) Both ends of P3, P8, Pn (n≥10) are adjacent to a unique support vertex u∈S(T). (c) ∀u∈S(T), except for its leaf, either u is adjacent to one of the endpoints of P3, P8, Pn (n≥10), or u is adjacent to other vertices in S(T). Then, γr(T)≠γ(T)+1.
In fact, if the tree T has and only has the path P3, then P3 is also star K1,2. According to Lemma 5, γr(T)=γ(T)≠γ(T)+1. If the tree T has and only has the path P8, as shown in Figure 6, based on assumptions, S(T)∪{v3,v6} can dominate the tree T, so γ(T)≤|S(T)∪{v3,v6}|=|S(T)|+2. v1 is adjacent to u1, v8 is adjacent to u2, and u1,u2∈S(T). From Lemma 4, γ(P6)=⌈63⌉=2. So, γ(T)≥|S(T)|+2. Therefore, γ(T)=|S(T)|+2. Since the leaves of T one-to-one correspond to their support vertices, and each leaf requires at least one legion for security defense or security defense by its support vertex, and if the vertex v1 for security defense, either f(v1)=1, or f(v2)=1, or f(u1)=2. The same if the vertex v8 for security defense, either f(v8)=1, or f(v7)=1, or f(u2)=2. That is, the vertex v1 cannot be securely defended by u1, and the vertex v8 cannot be securely defended by u2, unless u1 or u2 are stationed in two legions. From Lemma 2, γr(P8)=⌈3×87⌉=4, so γr(T)≥|S(T)|+4=γ(T)+2>γ(T)+1. If the tree T has and only has the path Pn (n≥10), the same is true for the path P8 above. γ(T)=|S(T)|+⌈n−23⌉, and from Lemma 2, γr(Pn)=⌈3n7⌉. Then, when n≥10, γr(T)≥|S(T)|+⌈3n7⌉=γ(T)+⌈3n7⌉−⌈n−23⌉≥γ(T)+2>γ(T)+1.
Theorem9. Suppose the tree T=(V,E) does not contain any strong support vertices, contains m (m is a positive integer) K2 (let ui,li∈V, uili∈E,i=1,2,⋯,m) and n (n is a positive integer) stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n) (the set formed by their center vertices is C, and the set formed by their outer vertices is W) and has only one of the following seven cases: (1) K1; (2) P2; (3) P4; (4) P5; (5) P6; (6) P7; (7) P9. Further, also suppose the following conditions are satisfied:
(a) Each K2 has a vertex ui∈S(T) and another vertex li∈L(T), S(T)={u1,u2,⋯,um}, and L(T)={l1,l2,⋯,lm}.
(b) ∀w∈W, there is a unique support vertex u∈S(T), such that wu∈E(T), and ∀v∈C, v is at least 2 away from every vertex in S(T).
(c) Both ends of K1, P2, P4, P5, P6, P7, P9 are adjacent to a unique support vertex u∈S(T).
(d) ∀u∈S(T), except for its leaf, either u is adjacent to an outer vertex w∈W, or u is adjacent to one of the endpoints of K1, P2, P4, P5, P6, P7, P9, or u is adjacent to other vertices in S(T).
As shown in Figure 7, then, γr(T)=γ(T)+1.
Proof. By Theorem 2, there is a γ(T)-set S such that S(T)⊆S. Then, divided into seven cases, the proof is as follows:
(1) If the tree T has only the K1, as shown in Figure 7(T1), then, obviously, S(T1)∪C(T1) is a dominating set of tree T1, and S(T1)∩C(T1)=∅, so γ(T1)≤|S(T1)∪C(T1)|=|S(T1)|+|C(T1)|. Based on assumptions, ∀ v∈C(T1), v is at least 2 away from every vertex in S(T1), so γ(T1)≥|S(T1)|+|C(T1)|. Hence, γ(T1)=|S(T1)|+|C(T1)|, and thus S(T1)∪C(T1) is a γ(T1)-set of the tree T1.
Let f=(V0,V1,V2)=(V−S(T1)∪C(T1)∪{v1},S(T1)∪C(T1)∪{v1},∅), and then it has no undefended vertex. It is clear that the vertices in V−S(T1)∪C(T1)∪{v1} consist of the set of all leaves L(T1) and the set of outer vertices W(T1) of the stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n), that is, V−S(T1)∪C(T1)∪{v1}=L(T1)∪W(T1), and L(T1)∩W(T1)=∅. ∀u∈V−S(T1)∪C(T1)∪{v1}=L(T1)∪W(T1), if u∈L(T1), since the tree T1 does not contain any strong support vertices, let uv∈E(T1) and v∈S(T1), and let fu=(V∗0,V∗1,V∗2)=(L(T1)∪W(T1)∪{v}∖{u},S(T1)∪C(T1)∪{v1,u}∖{v},∅). Then, obviously, fu has no undefended vertex. If u∈W(T1), let uv∈E(T1) and v∈C(T1), and let fu=(V∗0,V∗1,V∗2)=(L(T1)∪W(T1)∪{v}∖{u},S(T1)∪C(T1)∪{v1,u}∖{v},∅). Then, obviously, fu has no undefended vertex. So, f is a WRDF of the tree T1. Therefore, γr(T1)≤w(f)=|S(T1)∪C(T1)∪{v1}|=|S(T1)∪C(T1)|+1=|S(T1)|+|C(T1)|+1=γ(T1)+1.
The leaves of T1 one-to-one correspond to their support vertices, and each leaf requires at least one legion for security defense or security defense by its support vertex. ∀v∈C(T1), v is at least 2 away from every vertex in S(T1), so the vertex v requires at least one legion for security defense, and if the vertex v1 for security defense, either f(v1)=1, or f(u1)=2, or f(u2)=2, as shown in Figure 7(T1). So, γr(T1)≥|S(T1)∪C(T1)|+1=|S(T1)|+|C(T1)|+1=γ(T1)+1. Therefore, γr(T1)=γ(T1)+1.
(2) If the tree T has only the path P2, as shown in Figure 7(T2), the proof of (1) is the same.
(3) If the tree T has only the path P4, as shown in Figure 7(T3), based on assumptions, S(T3)∪C(T3)∪{v2} can dominate the tree T3, and S(T3)∩C(T3)=∅, so γ(T3)≤|S(T3)∪C(T3)∪{v2}|=|S(T3)|+|C(T3)|+1. Based on assumptions, ∀ v∈C(T3), v is at least 2 away from every vertex in S(T3). Since v1 is adjacent to u1, v4 is adjacent to u2 and u1,u2∈S(T3) and from Lemma 4, γ(P2)=⌈23⌉=1. So, γ(T3)≥|S(T3)|+|C(T3)|+1. Therefore, γ(T3)=|S(T3)|+|C(T3)|+1, and thus S(T3)∪C(T3)∪{v2} is a γ(T3)-set of the tree T3.
Let f=(V0,V1,V2)=(V−S(T3)∪C(T3)∪{v2,v3},S(T3)∪C(T3)∪{v2,v3},∅), and then it has no undefended vertex. Obviously, the vertices in V−S(T3)∪C(T3)∪{v2,v3} consist of three parts: (i) the set of all leaves L(T3); (ii) the set of outer vertices W(T3) of the stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n); (iii) the set {v1,v4}, that is, V−S(T3)∪C(T3)∪{v2,v3}=L(T3)∪W(T3)∪{v1,v4} and L(T3)∩W(T3)=∅, L(T3)∩{v1,v4}=∅, W(T3)∩{v1,v4}=∅. ∀u∈V−S(T3)∪C(T3)∪{v2,v3}=L(T3)∪W(T3)∪{v1,v4}, if u∈L(T3), since the tree T3 does not contain any strong support vertices, let uv∈E(T3) and v∈S(T3), and let fu=(V∗0,V∗1,V∗2)=(L(T3)∪W(T3)∪{v1,v4,v}∖{u},S(T3)∪C(T3)∪{v2,v3,u}∖{v},∅). Then, obviously, fu has no undefended vertex. If u∈W(T3), let uv∈E(T3) and v∈C(T3), and let fu=(V∗0,V∗1,V∗2)=(L(T3)∪W(T3)∪{v1,v4,v}∖{u},S(T3)∪C(T3)∪{v2,v3,u}∖{v},∅). Then, obviously, fu has no undefended vertex. If u=v1, let fu=(V∗0,V∗1,V∗2)=(L(T3)∪W(T3)∪{v2,v4},S(T3)∪C(T3)∪{v1,v3},∅), and then, obviously, fu has no undefended vertex. If u=v4, let fu=(V∗0,V∗1,V∗2)=(L(T3)∪W(T3)∪{v1,v3},S(T3)∪C(T3)∪{v2,v4},∅). Then, obviously, fu has no undefended vertex. So f is a WRDF of the tree T3. Therefore, γr(T3)≤w(f)=|S(T3)∪C(T3)∪{v2,v3}|=|S(T3)|+|C(T3)|+2=γ(T3)+1.
The leaves of T3 one-to-one correspond to their support vertices, and each leaf requires at least one legion for security defense or security defense by its support vertex. ∀v∈C(T3), v is at least 2 away from every vertex in S(T3), so the vertex v requires at least one legion for security defense, and if the vertex v1 for security defense, either f(v1)=1, or f(v2)=1, or f(u1)=2. The same if the vertex v4 for security defense, either f(v4)=1, or f(v3)=1, or f(u2)=2, as shown in Figure 7(T3). That is, the vertex v1 cannot be securely defended by u1, and the vertex v4 cannot be securely defended by u2, unless u1 or u2 are stationed in two legions. From Lemma 2, γr(P4)=⌈3×47⌉=2, so γr(T3)≥|S(T3)|+|C(T3)|+2=γ(T3)+1. Therefore, γr(T3)=γ(T3)+1.
(4) If the tree T has only the path P5, P6, P7, P9, as shown in Figure 7(T4, T5, T6, T7), the proof of (3) is the same.
Note2. Suppose the tree T=(V,E) does not contain any strong support vertices, contains m (m is a positive integer) K2 (let ui,li∈V, uili∈E,i=1,2,⋯,m) and n (n is a positive integer) stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n) (the set formed by their center vertices is C, and the set formed by their outer vertices is W), and has only one of the following three cases: (1) P3; (2) P8; (3) Pn (n≥10). Further suppose also that the following conditions are satisfied: (a) Each K2 has a vertex ui∈S(T) and another vertex li∈L(T), S(T)={u1,u2,⋯,um}, and L(T)={l1,l2,⋯,lm}. (b) ∀w∈W, there is a unique support vertex u∈S(T), such that wu∈E(T), and ∀v∈C, v is at least 2 away from every vertex in S(T). (c) Both ends of P3, P8, Pn (n≥10) are adjacent to a unique support vertex u∈S(T). (d) ∀u∈S(T), except for its leaf, either u is adjacent to an outer vertex w∈W, or u is adjacent to one of the endpoints of P3, P8, Pn (n≥10), or u is adjacent to other vertices in S(T). Then, γr(T)≠γ(T)+1.
In fact, if the tree T has only the path P3, then P3 is also star K1,2, and according to Lemma 5, γr(T)=γ(T)≠γ(T)+1. If the tree T has only the path P8, as shown in Figure 8, based on assumptions, S(T)∪C(T)∪{v3,v6} can dominate the tree T, and S(T)∩C(T)=∅, so γ(T)≤|S(T)∪C(T)∪{v3,v6}|=|S(T)|+|C(T)|+2. Based on assumptions, ∀ v∈C(T), v is at least 2 away from every vertex in S(T). v1 is adjacent to u1, and v8 is adjacent to u2, and u1,u2∈S(T). With regards from Lemma 4, γ(P6)=⌈63⌉=2. So, γ(T)≥|S(T)|+|C(T)|+2. Therefore, γ(T)=|S(T)|+|C(T)|+2. The leaves of T one-to-one correspond to their support vertices, and each leaf requires at least one legion for security defense or security defense by its support vertex. ∀v∈C(T), v is at least 2 away from every vertex in S(T), so the vertex v requires at least one legion for security defense, and if the vertex v1 for security defense, either f(v1)=1, or f(v2)=1, or f(u1)=2. The same if the vertex v8 for security defense, either f(v8)=1, or f(v7)=1, or f(u2)=2, as shown in Figure 8. That is, the vertex v1 cannot be securely defended by u1 and the vertex v8 cannot be securely defended by u2, unless u1 or u2 are stationed in two legions. From Lemma 2, γr(P8)=⌈3×87⌉=4, so γr(T)≥|S(T)|+|C(T)|+4=γ(T)+2>γ(T)+1. If the tree T has only the path Pn (n≥10), the same is true for the path P8 above, γ(T)=|S(T)|+|C(T)|+⌈n−23⌉, and from Lemma 2, γr(Pn)=⌈3n7⌉. Then, when n≥10, γr(T)≥|S(T)|+|C(T)|+⌈3n7⌉=γ(T)+⌈3n7⌉−⌈n−23⌉≥γ(T)+2>γ(T)+1.
Theorem10. Suppose the tree T=(V,E) does not contain any strong support vertices, contains m (m is a positive integer) K2 (let ui,li∈V, uili∈E,i=1,2,⋯,m) and n (n is a positive integer) stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n) and has only one connected branch from star K∗1,t (t≥2) and an outer vertex w∗ of star K∗1,t (t≥2) connected to a vertex v∗ (the set formed by the center vertices of all stars is C, and the set formed by the outer vertices is W). Further, suppose the following conditions are satisfied:
(a) Each K2 has a vertex ui∈S(T) and another vertex li∈L(T), S(T)={u1,u2,⋯,um}, and L(T)={l1,l2,⋯,lm}.
(b) ∀w∈W∖{w∗}, there is a unique support vertex u∈S(T), such that wu∈E(T), and ∀v∈C, v is at least 2 away from every vertex in S(T).
(c) ∃u∗∈S(T), such that u∗v∗∈E(T).
(d) ∀u∈S(T)∖{u∗}, except for its leaf, either u is adjacent to an outer vertex w∈W∖{w∗}, or u is adjacent to other vertices in S(T).
As shown in Figure 9, then, γr(T)=γ(T)+1.
Proof. By Theorem 2, there is a γ(T)-set S such that S(T)⊆S. Based on assumptions, the set formed by the center vertices of the stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n) and the star K∗1,t (t≥2) is C(T). Then, obviously, S(T)∪C(T) is a dominating set of tree T, and S(T)∩C(T)=∅, so γ(T)≤|S(T)∪C(T)|=|S(T)|+|C(T)|. Based on assumptions, ∀ v∈C(T), v is at least 2 away from every vertex in S(T), so γ(T)≥|S(T)|+|C(T)|. Hence, γ(T)=|S(T)|+|C(T)|, and thus S(T)∪C(T) is a γ(T)-set of the tree T.
Let f=(V0,V1,V2)=(V−S(T)∪C(T)∪{v∗},S(T)∪C(T)∪{v∗},∅), and then it has no undefended vertex, as shown in Figure 9. Obviously, the vertices in V−S(T)∪C(T)∪{v∗} consist of two parts: (i) the set of all leaves L(T); (ii) the set of outer vertices W(T) of the stars K1,t1,K1,t2,⋯,K1,tn(ti≥2,i=1,2,⋯,n) and the star K∗1,t (t≥2), that is, u∈V−S(T)∪C(T)∪{v∗}=L(T)∪W(T) and L(T)∩W(T)=∅. ∀u∈V−S(T)∪C(T)∪{v∗}=L(T)∪W(T), if u∈L(T), since the tree T does not contain any strong support vertices, let uv∈E(T) and v∈S(T), and let fu=(V∗0,V∗1,V∗2)=(L(T)∪W(T)∪{v}∖{u},S(T)∪C(T)∪{v∗,u}∖{v},∅). Then, obviously, fu has no undefended vertex. If u∈W(T), let uv∈E(T) and v∈C(T), and let fu=(V∗0,V∗1,V∗2)=(L(T)∪W(T)∪{v}∖{u},S(T)∪C(T)∪{v∗,u}∖{v},∅). Then, obviously, fu has no undefended vertex. So, f is a WRDF of the tree T. Therefore, γr(T)≤w(f)=|S(T)∪C(T)∪{v∗}|=|S(T)|+|C(T)|+1=γ(T)+1.
The leaves of T one-to-one correspond to their support vertices, and each leaf requires at least one legion for security defense or security defense by its support vertex. ∀v∈C(T), v is at least 2 away from every vertex in S(T), so the vertex v requires at least one legion for security defense, and if the vertex v∗ for security defense, either f(v∗)=1, or f(w∗)=1, or f(u∗)=2, as shown in Figure 9, so γr(T)≥|S(T)∪C(T)|+1=|S(T)|+|C(T)|+1=γ(T)+1. Therefore, γr(T)=γ(T)+1.
In this paper, we give some sufficient conditions for a tree to have its weak Roman domination number be equal to its domination number plus 1 (γr(T)=γ(T)+1) by recursion and construction, such as Theorems 6–10. Conversely, we hold that the trees T satisfying γr(T)=γ(T)+1 only have the characteristics of Theorems 6–10, which is subject to further verification. We will further investigate some characteristics of trees in which the weak Roman domination number is equal to the domination number plus k (γr(T)=γ(T)+k).
The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.
We thank the associate editor and the reviewers for their useful feedback that improved this paper.
This work is supported partly by the National Natural Science Foundation (NNSF) of China under Grants 62073122 and 61203050, the Plan of Key Scientific Research Projects of Colleges and Universities in Henan Province (No. 22A880007), the National Natural Science Foundation of Henan (No. 202300410343), the Postgraduate Education Reform and Quality Improvement Project of Henan Province (No. YJS2022ZX34), the Soft Science Research Project of 2023 Science and Technology Development Plan of Henan Province (No. 232400410210) and the Higher Education Teaching Reform research and practice project in Henan Province (No. 2019SJGLX735).
We declare no conflict of interest.
[1] |
I. Stewart, Defend the Roman Empire, Sci. Am., 281 (1999), 136–138. https://doi.org/10.1038/scientificamerican1299-136 doi: 10.1038/scientificamerican1299-136
![]() |
[2] |
E. J. Cockayne, P. A. Jr. Dreyer, S. M. Hedetniemi, S. T. Hedetniemi, Roman domination in graphs, Disc. Math., 278 (2004), 11–22. https://doi.org/10.1016/j.disc.2003.06.004 doi: 10.1016/j.disc.2003.06.004
![]() |
[3] |
A. Klobučar, I. Puljić, Roman domination number on cardinal product of paths and cycles, Croat. Oper. Res. Rev., 6 (2015), 71–78. https://doi.org/10.17535/crorr.2015.0006 doi: 10.17535/crorr.2015.0006
![]() |
[4] |
M. A. Henning, S. T. Hedetniemi, Defending the Roman Empire-A new strategy, Discrete Math., 266 (2003), 239–251. https://doi.org/10.1016/S0012-365X(02)00811-7 doi: 10.1016/S0012-365X(02)00811-7
![]() |
[5] |
R. Hernández-Ortiz, L. P. Montejano, J. A. Rodríguez-Velázquez, Weak Roman domination in rooted product graphs, AIMS Math., 6 (2021), 3641–3653. https://doi.org/10.3934/math.2021217 doi: 10.3934/math.2021217
![]() |
[6] | J. Yang, J. L. Song, Trees T with weak Roman domination number being equal to domination number, Mathematics in Practice and Theory (in Chinese), 43 (2013), 134–140. |
[7] | J. Yang, Z. Q. Li, Some properties of weak Roman graph and weak Roman domination in graphs, Chin. J. of Eng. Math. (in Chinese), 39 (2022), 621–630. |
[8] |
F. N. Pour, H. A. Ahangar, M. Chellali, S. M. Sheikholeslami, Global triple Roman dominating function, Discrete Appl. Math., 314 (2022), 228–237. https://doi.org/10.1016/j.dam.2022.02.015 doi: 10.1016/j.dam.2022.02.015
![]() |
[9] |
Z. Li, A note on the bounds of Roman domination numbers, AIMS Math., 6 (2021), 3940–3946. https://doi.org/10.3934/math.2021234 doi: 10.3934/math.2021234
![]() |
[10] |
M. Chellali, T. W. Haynes, S. T. Hedetniemi, Roman and total domination, Quaest. Math., 38 (2015), 749–757. https://doi.org/10.2989/16073606.2015.1015647 doi: 10.2989/16073606.2015.1015647
![]() |
[11] |
Z. Du, A. A. S. A. Jamri, R. Hasni, D. A. Mojdeh, Maximal first Zagreb index of trees with given Roman domination number, AIMS Math., 7 (2022), 11801–11812. https://doi.org/10.3934/math.2022658 doi: 10.3934/math.2022658
![]() |
[12] |
S. C. García, A. C. Martínez, I. G. Yero, Quasi-total Roman domination in graphs, Results Math., 74 (2019), 173. https://doi.org/10.1007/s00025-019-1097-5 doi: 10.1007/s00025-019-1097-5
![]() |
[13] |
P. R. L. Pushpam, S. Padmapriea, Global Roman domination in graphs, Discrete Appl. Math., 200 (2016), 176–185. https://doi.org/10.1016/j.dam.2015.07.014 doi: 10.1016/j.dam.2015.07.014
![]() |
[14] |
H. A. Ahangar, M. A. Henning, V. Samodivkin, I. G. Yero, Total Roman domination in graphs, Appl. Anal. Discr. Math., 10 (2016), 501–517. https://doi.org/10.2298/AADM160802017A doi: 10.2298/AADM160802017A
![]() |
[15] |
J. Amjadi, S. M. Sheikholeslami, M. Soroudi, Nordhaus-Gaddum bounds for total Roman domination, J. Comb. Optim., 35 (2018), 126–133. https://doi.org/10.1007/s10878-017-0158-5 doi: 10.1007/s10878-017-0158-5
![]() |
[16] | G. Gunther, B. L. Hartnell, L. Markus, D. Rall, Graphs with unique minimum domination sets, Congressus Numerantium, 101 (1994), 55–63. |
[17] |
Z. Gu, J. X. Meng, Z. Zhang, J. E. Wan, Some upper bounds related with domination number, J. Oper. Res. Soc. China, 1 (2013), 217–225. https://doi.org/10.1007/s40305-013-0012-0 doi: 10.1007/s40305-013-0012-0
![]() |
1. | Chang-Xu Zhang, Fu-Tao Hu, Shu-Cheng Yang, On the (total) Roman domination in Latin square graphs, 2024, 9, 2473-6988, 594, 10.3934/math.2024031 |