
Let f be a proper total k-coloring of a simple graph G from V(G)∪E(G) to {1,2,…,k}, let C(u,f) be the set of the colors assigned to the edges incident with u, and let nd(G) and Δ(G) denote the number of all vertices of degree d and the maximum degree in G, respectively. We call f a (2)-vertex distinguishing total k-coloring (k-(2)-vdc for short) if C(u,f)≠C(v,f) and C(u,f)∪{f(u)}≠C(v,f)∪{f(v)} for distinct vertices u,v∈V(G). The minimum number k of colors required for which G admits a k-(2)-vdc is denoted by χ″2s(G). In this paper, we show that a tree T with n2(T)≤n1(T) has χ″2s(T)=n1(T) if and only if T is not a tree with D(T)=2,3 or n1(T)=Δ(T), where D(T) is the diameter of tree T.
Citation: Chao Yang, Bing Yao, Zhi-xiang Yin. A new vertex distinguishing total coloring of trees[J]. AIMS Mathematics, 2021, 6(9): 9468-9475. doi: 10.3934/math.2021550
[1] | Ningge Huang, Lily Chen . AVD edge-colorings of cubic Halin graphs. AIMS Mathematics, 2023, 8(11): 27820-27839. doi: 10.3934/math.20231423 |
[2] | Baolin Ma, Chao Yang . Distinguishing colorings of graphs and their subgraphs. AIMS Mathematics, 2023, 8(11): 26561-26573. doi: 10.3934/math.20231357 |
[3] | Yinwan Cheng, Chao Yang, Bing Yao, Yaqin Luo . Neighbor full sum distinguishing total coloring of Halin graphs. AIMS Mathematics, 2022, 7(4): 6959-6970. doi: 10.3934/math.2022386 |
[4] | Gaixiang Cai, Fengru Xiao, Guidong Yu . The identification numbers of lollipop graphs. AIMS Mathematics, 2025, 10(4): 7813-7827. doi: 10.3934/math.2025358 |
[5] | Fugang Chao, Donghan Zhang . Neighbor sum distinguishing total choice number of IC-planar graphs with restrictive conditions. AIMS Mathematics, 2023, 8(6): 13637-13646. doi: 10.3934/math.2023692 |
[6] | Meiqin Jin, Ping Chen, Shuangliang Tian . Interval edge colorings of the generalized lexicographic product of some graphs. AIMS Mathematics, 2024, 9(11): 30597-30611. doi: 10.3934/math.20241477 |
[7] | Xiaoxue Hu, Jiangxu Kong . An improved upper bound for the dynamic list coloring of 1-planar graphs. AIMS Mathematics, 2022, 7(5): 7337-7348. doi: 10.3934/math.2022409 |
[8] | Zhenzhen Jiang, Jun Yue, Xia Zhang . Polychromatic colorings of hypergraphs with high balance. AIMS Mathematics, 2020, 5(4): 3010-3018. doi: 10.3934/math.2020195 |
[9] | Miao Fu, Yuqin Zhang . Results on monochromatic vertex disconnection of graphs. AIMS Mathematics, 2023, 8(6): 13219-13240. doi: 10.3934/math.2023668 |
[10] | Jing Su, Qiyue Zhang, Bing Yao . The connection between the magical coloring of trees. AIMS Mathematics, 2024, 9(10): 27896-27907. doi: 10.3934/math.20241354 |
Let f be a proper total k-coloring of a simple graph G from V(G)∪E(G) to {1,2,…,k}, let C(u,f) be the set of the colors assigned to the edges incident with u, and let nd(G) and Δ(G) denote the number of all vertices of degree d and the maximum degree in G, respectively. We call f a (2)-vertex distinguishing total k-coloring (k-(2)-vdc for short) if C(u,f)≠C(v,f) and C(u,f)∪{f(u)}≠C(v,f)∪{f(v)} for distinct vertices u,v∈V(G). The minimum number k of colors required for which G admits a k-(2)-vdc is denoted by χ″2s(G). In this paper, we show that a tree T with n2(T)≤n1(T) has χ″2s(T)=n1(T) if and only if T is not a tree with D(T)=2,3 or n1(T)=Δ(T), where D(T) is the diameter of tree T.
The problem of vertex distinguishing colorings can be traced to two articles [3] and [4]. Burris and Schelp [3] introduced a proper edge k-coloring of a simple graph G is called a vertex distinguishing edge k-coloring (k-vdec, or vdec for short) if for any two distinct vertices u and v of G, the set of the colors assigned to the edges incident with u differs from the set of the colors assigned to the edges incident with v. The minimum number of colors required for all vertex distinguishing colorings of G is denoted by χ′s(G). Let nd(G) denote the number of all vertices of degree d in G, or nd=nd(G) if no confusion. Furthermore, Burris and Schelp presented the following conjecture: Let G be a simple graph having no isolated edges and at most one isolated vertex, and let k be the smallest integer such that (kd)≥nd for all d with respect to δ(G)≤d≤Δ(G). Then k≤χ′s(G)≤k+1. The above Conjecture is known for some families of graphs, including complete graphs, complete bipartite graphs and many trees [3]. Later it has been proved for graphs of large maximum degree [1] and for G a union of cycles or a union of paths [2]. Therefore, the most challenging cases seem to occur when G has small maximum degree which is at least three. In this paper we show some results on trees to confirm positively Burris-Schelp Conjecture.
Graphs mentioned here are undirected, finite and simple, and graph colorings are proper (so we omit the term "proper" hereafter). We use standard terminology and notation of graph theory, and introduce a new total coloring with two vertex distinguishing constraints. Let f be a total coloring of a simple graph G from V(G)∪E(G) to {1,2,…,k}, and let C(u,f) be the set of the colors assigned to the edges incident with u. If f holds simultaneously
(DC):C(w,f)≠C(z,f)andC(w,f)∪{f(w)}≠C(z,f)∪{f(z)}for distinctw,z∈V(G), |
then we call call f a (2)-vertex distinguishing total k-coloring, and write k-(2)-vdc for short. The minimum number k of colors required for which G admits a k-(2)-vdc is denoted by χ″2s(G). Clearly, an edge coloring obtained by removing the colors assigned to vertices from a k-(2)-vdc of G is just a k-vdec of G, which implies that χ′s(G)≤χ″2s(G). We say a color a is missing at a vertex u means a∉C(u,f)∪{f(u)}. A leaf is a vertex of degree one, and a k-degree vertex is one of degree k at least two. The set of neighbors of a vertex u of G is denoted by NG(u), or N(u) if no confusion; the degree dG(u) of u is equal to |NG(u)|. D(G) stands for the diameter of G. The shorthand notation [m,n] stands for an integer set {m,m+1,…,n} with n>m≥0. Let Pn be a path on n vertices. In the present work, we will show the following:
Theorem 1. Let T be a tree with n2(T)≤n1(T) and |T|≥3. Then χ″2s(T)=n1(T)+2 if T≅P3,P4; χ″2s(T)=n1(T)+1 if T≇P3,P4 and n1(T)=Δ(T) or D(T)=3; and χ″2s(T)=n1(T) for otherwise.
Lemma 2. Let T be a tree with n2(T)≥n1(T). Then there exists a 2-degree vertex such that one of its neighbors is either a leaf or a vertex of degree 2.
Proof. Suppose (reductio and absurdum) that each 2-degree vertex x is not adjacent to a leaf and has its neighborhood N(x)={x1,x2} such that dT(xi)≥3 for i=1,2. Then we have n2≤Δ∑d=3nd by the assumption. Applying the formula n1=2+Δ∑d=3(d−2)nd shown in [5], we have n1≥2+nd for d≥3, and
n1≥2+n3+2Δ∑d=4nd=2+2Δ∑d=3nd−n3, |
which shows that Δ∑d=3nd≤12(n1+n3)−1. It immediately deduces that
n2≤Δ∑d=3nd≤12(n1+n3)−1≤n1−2, |
which contradicts to the hypothesis.
Lemma 3. Let T be a tree with D(T)=3. Then χ″2s(T)=n1(T)+2 for |V(T)|=4, and χ″2s(T)=n1(T)+1 for |V(T)|≥5.
Proof. Observe that each tree of diameter 3, also, is a double star DSm+1,n+1 obtained by joining two centers of two complete bipartite graphs K1,m and K1,n. When n=m=1, T is a path of length 3 with n1(T)=n2(T), and we can verify χ″2s(T)=n1(T)+2 by simply assigning colors to the edges and vertices of T. For m+n≥3, let V(K1,m)={s,s1,s2,…,sm} and E(K1,m)={ss1,ss2,…,ssm}; V(K1,n)={t,t1,t2,…,tn} and E(K1,n)={tt1,tt2,…,ttn}, where s and t are the centers. Then V(DSm+1,n+1)=V(K1,m)∪V(K1,n) and E(DSm+1,n+1)=E(K1,m)∪E(K1,n)∪{st}. Suppose that f is a k-(2)-vdc of DSm+1,n+1 such that k=χ″2s(DSm+1,n+1). Since C(si,f)≠C(sj,f) and C(ti,f)≠C(tj,f) for i≠j, C(si,f)≠C(tl,f) for i≠l, and C(s,f)≠C(t,f), k≥m+n+1. This k-(2)-vdc f can be exactly defined by setting f(ssi)=i for i∈[1,m], and f(si)=i−1 for i∈[2,m], f(s1)=m; f(s)=m+2, f(st)=m+1, f(t)=1; f(ttj)=m+1+j for i∈[1,n], and f(tj)=m+j for i∈[2,n], f(t1)=m+n+1.
Lemma 4. For n≥2, a 2-star, denoted as K(2)1,2n, is a tree on 1+2n vertices such that every leaf of K1,n is joined a new vertex out of K1,n by an edge. Then χ″2s(K(2)1,2n)=n+1.
Proof. Suppose that K1,n has its own vertex set V(K1,n)={w,w1,w2,…,wn} and edge set E(K1,n)={ww1,ww2,…,wwn}. By the definition of a 2-star K(2)1,2n, we have V(K(2)1,2n)=V(K1,n)∪S, where S={z1, z2, …, zn} with S∩V(K1,n)=∅, and edge set E(K(2)1,2n)=E(K1,n)∪{w1z1, w2z2, …, wnzn}. Clearly, n1(K(2)1,2n)=n, each zi is a leaf of K(2)1,2n for i∈[1,n]. It is clear that χ″2s(K(2)1,2n)≥n+1. It is straightforward to define an (n+1)-(2)-vdc f of K(2)1,2n as follows: f(w)=n+1; f(wwi)=i for i∈[1,n]; f(wizi)=i−1 for i∈[2,n], and f(w1z1)=n; f(wi)=i+1 for i∈[1,n−1], and f(wn)=1; f(zi)=i for i∈[3,n], and f(z1)=f(z2)=n+1.
Lemma 5. Let T be a tree with n2(T)≤n1(T) and |T|≥3. If Δ(T)=n1(T), then χ″2s(T)=n1(T)+2 for Δ(T)=2, and χ″2s(T)=n1(T)+1 for Δ(T)≥3.
Proof. Write ni=ni(T) for i∈[δ,Δ], where δ=δ(T) and Δ=Δ(T). By a formula n1=2+∑3≤d≤Δ(d−2)nd in [5] and Δ=n1, we have nΔ=1 and nd=0 for 3≤d≤Δ−1. Thus, T is called a spider having its body w with dT(w)=Δ and legs Pi=wwi,1wi,2⋯wi,ki for ki≥1 and i∈[1,Δ]. When ki=2 for i∈[1,Δ], T=K(2)1,2Δ is a 2-star. When Δ(T)=2, T is a path of length 2 or 3, hence it easily gets that χ″2s(T)=n1+2. For Δ(T)≥3, we show χ″2s(T)=n1(T)+1 in the following. If T is a 2-star, we are done by Lemma 4. By induction on the order of spiders, and assume that T is not a 2-star.
Observe that there exists some ki=1 since T is not a 2-star and n2(T)≤n1(T). Without loss of generality, set k1=1, then we have a tree T1=T−w1,1 with n1(T1)=n1−1 and Δ(T1)=Δ−1.
When n2(T1)≤n1(T1), by induction hypothesis, T1 has a p1-(2)-vdc f with χ″2s(T1)=p1=n1(T1)+1. Then we can extend f to an (n1+1)-(2)-vdc g of T as follows: g(z)=f(z) for z∈(V(T)∖{w1,1})∪(E(T)∖{ww1,1}), g(ww1,1)=p1+1, g(w1,1)∈[1,p1]∖{f(w)}.
When n1(T1)=n2(T1)−1=n2−1, we meet every kj≤2 for j∈[2,Δ]. In this case, T1 is a spider with χ″2s(T1)=n1(T1)+1=n1, the lemma holds. For some ki≥3 with ki≠k1, we can take a tree T2=T1−wi,ki−1+wi,ki−2wi,ki, here wi,ki−2≠w. Clearly, n1(T2)=n2(T2)=n2−1, so T1 has a p2-(2)-vdc f′ with χ″2s(T2)=p2=n1(T2)+1=n1(T1)+1=n1 by induction hypothesis. We define a (n1+1)-(2)-vdc g′ of T by setting g′(z)=f′(z) for z∈(V(T)∖{w1,1,wi,ki−1})∪(E(T)∖{ww1,1,wi,ki−2wi,ki−1,wi,ki−1wi,ki}); g′(ww1,1)=p2+1, g′(w1,1)∈[1,p2]∖{f′(w)}; g′(wi,ki−2wi,ki−1)=p2+1, g′(wi,ki−1)∈[1,p2]∖{f′(wi,ki−2),f′(wi,ki−2wi,ki),f′(wi,ki)}; and g′(wi,ki−1wi,ki)=f′(wi,ki−2wi,ki).
This completes the proof.
Let ni=ni(T) for i=1,2, and N(w)=NT(w) for w∈V(T). If T is a path of length 2 or 3, then it is not hard to prove that χ″(2)s(T)=n1+2. According to Lemmas 3–5, we will show that χ″(2)s(T)=n1 by induction on the order of trees. Assume that T is not a 2-star, and suppose that n2≤n1, Δ(T)<n1 and D(T)≥5 in the following discussion.
Case A. There exists a leaf v having a neighbor u with dT(u)≥4. Let T′=T−v. Then n1(T′)=n1−1, n2(T′)=n2 and D(T′)=D(T)≥5. If Δ(T′)=n1(T′), then T′ is a spider, which means that dT′(u)=Δ(T′), and hence Δ(T)=n1, a contradiction. Hence, Δ(T′)<n1(T′).
Case A.1. n1(T′)≥n2(T′). By induction hypothesis, there is a total coloring θT′:V(T′)∪E(T′)→[1,b′] such that χ″2s(T′)=b′=n1(T′)=n1−1. It is straightforward to define an n1-(2)-vdc θ:V(T)∪E(T)→[1,b′]∪{b′+1} by setting θ(z)=θT′(z) for z∈(V(T)∪E(T))∖{uv,v}, θ(uv)=b′+1 and θ(v)=a, where the color a∈[1,b′] is missing at u.
Case A.2. n1(T′)=n2(T′)−1. By Lemma 2, T′ has a 2-degree vertex x with its neighborhood NT′(x)={x1,x2} having dT′(x1)≤2. Note that dT′(x)=dT(x) and dT′(x1)=dT(x1). We get a tree T1=T′−x+x1x2. Clearly, n1(T1)=n1(T′)>Δ(T′)=Δ(T1), n2(T1)=n2(T′)−1 and D(T1)≥D(T′)−1=D(T)−1. If D(T1)=4, and n2(T1)=n2(T′)−1=n1(T′)=n1(T1), then we have n1(T1)=Δ(T1)=Δ(T′)<n1(T1), a contradiction. Hence D(T1)≥5. By induction hypothesis, T1 admits a (2)-vdc θT1:V(T1)∪E(T1)→[1,b1] such that χ″2s(T1)=b1=n1(T1)=n1−1. We define a total coloring θ of T as: θ(z)=θT1(z) for z∈V(T)∪E(T)∖{v,uv,x,xx1,xx2}; θ(uv)=b1+1 and θ(v)=a, where the color a∈[1,b1] is missing at u; θ(xx2)=b1+1, θ(xx1)=θT1(x1x2), and θ(x)∈[1,b1]∖{θT1(x1x2),θT1(x1),θT1(x2)}. It is not hard to check that θ holds (DC). Therefore, θ is a desired n1-(2)-vdc.
Case B. There is a leaf v having a 3-degree neighbor u in T, and Case A does not appear.
Case B.1. v′ is another leaf in the neighborhood N(u)={v,v′,u′}, and T has a 2-degree vertex x having its neighborhood {x1,x2} such that x1 is a leaf of T.
We have a tree T1=T−{v,v′,x}+x1x2. Clearly, n1(T1)=n1−1 and n2(T1)=n2−1. We shall frist discuss two subcases D(T1)=3 or 4 as below.
Case B.1.1. If D(T1)=3, then T1 is one of two trees shown in Figure 1(a) and 1(b). So, n1=5, D(T)=5. Then T admits a 5-(2)-vdc shown in Figure 1(c).
Case B.1.2. If D(T1)=4, let w0 be the center of T1. We consider dT1(w0)=2. So, T1 is one of Figure 2(a), 2(b) and 2(c), which implies the structural of T. To determine χ″2s(T), we give a 5-(2)-vdc of T by the graph shown in Figure 2(d).
When D(T1)=4 and dT1(w0)≥3, T is as one taken from Figure 3(a) and 3(b). We take a vertex w′∈N(w0)∖{u,x2} such that dT1(w′)≤dT1(z) for z∈N(w0)∖{u,x2}.
If dT1(w′)=3 (see Figure 3(a) or 3(b)), we take another tree T′1=T−{w′,w1,w2}, where w1,w2∈N(w′). Notice that n1(T′1)=n1−2, n2(T′1)=n2, D(T′1)=D(T) and n1(T′1)>Δ(T′1)≥n2(T′1). By induction hypothesis, we get a (2)-vdc θT′1:V(T′1)∪E(T′1)→[1,b′1] such that χ″(2)s(T′1)=b′1=n1(T′1). We can extend the total coloring θT′1 to an n1-(2)-vdc θ of T as follows: θ(z)=θT′1(z) for z∈V(T)∪E(T)∖{uv,w′,w1,w2,w0w′,w′w1,w′w2}; θ(uv)=b′1+1; and θ(w0w′)=b′1+1, θ(w′w1)=θT′1(uv), θ(w1)=θ(w2)=θT′1(v), θ(w′w2)=b′1+2, and θ(w′)∈[1,b′1]∖{θT′1(v),θT′1(w0),θT′1(uv)}.
If dT1(w′)=2 (see Figure 3(c)), then we get a tree T′2=T−{w′,w″}, where w″∈N(w′)={w0,w″}. Hence, n1(T′2)=n1−1, n2(T′2)=n2−1, D(T′2)=D(T), n1(T′2)>Δ(T′2)≥n2(T′2). By induction hypothesis, there is a (2)-vdc θT′2:V(T′2)∪E(T′2)→[1,b′2] such that χ″2s(T′2)=b′2=n1(T′2). Thereby, we define an n1-(2)-vdc θ of T in the way that θ(z)=θT′2(z) for z∈V(T)∪E(T)∖{uv,w′,w″,w0w′,w′w″}; θ(uv)=b′2+1; and θ(w0w′)=b′2+1, θ(w′w″)=θT′2(uv), θ(w″)=θT′2(v), and θ(w′)∈[1,b′2]∖{θT′2(v),θT′2(w0),θT′2(uv)}.
Case B.1.3. When D(T1)≥5, by induction hypothesis, T1 has a (2)-vdc θT1:V(T1)∪E(T1)→S1=[1,b1] such that χ″2s(T1)=b1=n1(T1)=n1−1. Notice that u is a leaf of T1. So, we can extend θT1 to a total coloring θ of T as follows: θ(z)=θT1(z) for z∈V(T)∪E(T)∖{v,v′,u,uv,uv′,x,xx1,xx2}, and θ(uu′)=θT1(u′), θ(u′)=b1+1, θ(uv′)=θT1(uu′), θ(v′)=θ(v)=θT1(u), θ(uv)=b1+1, θ(u)∈S1∖{θT1(u),θT1(u′),θT1(uu′)}; θ(xx1)=θT1(x1x2), and θ(xx2)=b1+1; θ(x)∈S1∖{θT1(x1),θT1(x1x2),θT1(x2)} such that C(x,θ)∪{θ(x)}≠C(u′,θ)∪{θ(u′)}. Hence, according to (DC), θ is a desired n1-(2)-vdc of T.
Case B.2. T does not have a 2-degree vertex being adjacent to a leaf of T. Let P=w1w2⋯wm−1wm be a longest path of T, where m is the diameter D(T) of T. So, dT(w2)=3=dT(wm−1) by the hypothesis of Case B, thus, w2 is adjacent to two leaves w1,w′1, and wm−1 is adjacent to two leaves wm,w′m.
Case B.2.1. If n2=0, then we take a tree H1=T−w1, and we get n1(H1)=n1−1 and n2(H1)=n2+1=1. Notice that D(H1)=D(T), n1(H1)≥3>n2(H1). By induction hypothesis, H1 admits a 2-vdc θH1:V(H1)∪E(H1)→[1,a1] such that χ″2s(H1)=a1=n1(H1). We define an n1-(2)-vdc θ of T as follows: θ(z)=θH1(z) for z∈V(T)∪E(T)∖{w1,w1w2}, and θ(w1w2)=a1+1, θ(w1)∈[1,a1]∖{θH1(w2)}.
Case B.2.2. When n2≥1 and dT(w3)=2, we take a tree H2=T−{w1,w′1,w2}. Notice that w3 is a leaf of H2, n1(H2)=n1−1, n2(H2)=n2−1. If D(H2)=3, then T is shown in Figure 4(a). It is easy to show that χ″2s(T)=n1 by assigning appropriately colors to the vertices and edges of T. If D(H2)=4, then T is one of Figure 4(b) and 4(c). Without loss of generality, T is Figure 4(b), so we can select a tree H3=T−w6 such that D(H3)=D(T), n1(H3)=n1−1≥3, n2(H3)=n2+1=2. Thereby, H3 has a (2)-vdc θH3:V(H3)∪E(H3)→[1,a3] such that χ″(2)s(H3)=a3=n1(H3). We define an n1-(2)-vdc θ of T as follows: θ(z)=θH3(z) for z∈V(T)∪E(T)∖{w6,w5w6}, and θ(w5w6)=a3+1, θ(w6)∈[1,a3]∖{θH3(w5)}.
Case B.2.3. When n2≥1 and dT(w3)≥3, let x be a 2-degree vertex of T having its own neighborhood N(x)={x1,x2} with dT(xi)≥2 for i=1,2. Consider a tree H4=T−{w1,w′1,x}+x1x2. Clearly, D(H4)≥D(T)−2≥3. If D(H4)=3, then T must be one shown in Figure 5(a), hence χ″2s(T)=n1 by assigning appropriately colors to the vertices and edges of T.
If D(H4)=4, then T is one of trees shown in Figure 5(b) and 5(c), say, T is shown in Figure 5(b). We have a tree H5=T−w1 such that D(H5)=D(T), n1(H5)=n1−1≥3, n2(H5)=n2+1=2. By induction hypothesis, there is a (2)-vdc θH5:V(H5)∪E(H5)→[1,a5] with χ″2s(H5)=a5=n1(H5)=n1−1, and we can extend it to an n1-(2)-vdc θ of T by setting θ(z)=θH5(z) for z∈V(T)∪E(T)∖{w1,w1w2}, and θ(w1w2)=a5+1, θ(w1)∈[1,a5]∖{θH5(w2)}.
Case C. Both Case A and Case B are false. In other words, each leaf is adjacent to a 2-degree vertex in T. Thereby, n1=n2.
Case C.1. Each 2-degree vertex x has its neighborhood N(x)={x1,x2} such that dT(x1)=1 and dT(x2)≥3, and no two 2-degree vertices have a common neighbor. Let Q=p1p2⋯pm be a longest path in T with m=D(T). Then dT(p3)≥3 and dT(pm−2)≥3 since n2=n1. Notice that dT(x)≤2 for x∈N(p3)∖{p2,p4} (resp. x∈N(pm−2)∖{pm−3,pm−1}) since x∉V(Q), and furthermore this vertex x is neither a leaf (Case B has been assumed to appear) nor a 2-degree vertex (by the assumption of this subcase and n2=n1). We confirm the latter subcase does not exist in the following.
Case C.2. Since Case C.1 does not exist, n2=n1, n1>Δ(T) and D(T)≥5, there exists a subgraph H having V(H)={w,u}∪V′ with r≥2, V′={yi,xi:i∈[1,r]} and E(H)={wu,uyi,yixi:i∈[1,r]}, where dT(w)≥3, moreover, every yi is a 2-degree vertex and every xi is a leaf for i∈[1,r]. Then we consider a tree T1=T−V′.
If D(T1)=1, then w is a leaf of T which contradicts to the hypothesis of Case C. If D(T1)=2, then T is a 2-star, which conflicts with D(T)≥5. If D(T1)=3, then T1 is a path of length 3, and hence n1=Δ(T), a contradiction. If D(T1)=4, then dT(w)=2 and n1=n2+1, a contradiction.
For D(T1)≥5, notice that n1(T1)=n1−r+1 and n2(T1)=n2−r, and u is a leaf of T1. By induction hypothesis, T1 admits a (2)-vdc θT1:V(T1)∪E(T1)→S∗=[1,b1] having χ″2s(T1)=b1=n1(T1)=n1−(r−1). We define a (2)-vdc θ of T as follows:
θ(z)=θT1(z) for z∈V(T1)∪E(T1)∖{u,uw};
θ(uyi)=b1+i for i∈[1,r−1], and θ(uyr)=θT1(uw);
θ(yjxj)=b1+j−1 for j∈[2,r], and θ(y1x1)=θT1(uw);
θ(w)=b1+1, θ(uw)=θT1(w); θ(xi)=θT1(u) for i∈[1,r]; and
θ(u)=a∈S∗∖{θT1(w),θT1(uw)}; θ(yi)=a′∈S∗∖{θT1(u),θT1(uw),a} for i∈[1,r].
Therefore, Theorem 1 follows from the principle of induction.
Corollary 6. Let G be a connected graph having cycles, p vertices and q edges. If n2(G)≤2(q−p+1)+n1(G), then χ″2s(G)≤1+2(q−p+1)+n1(G).
Proof. Let H be a spanning tree of G. Then we can construct another tree T of G by deleting each edge uv∈E′=E(G)∖E(H), and adding two new vertices u′,v′ by joining u′ with u and v′ with v, respectively. Clearly, n1(T)=2(q−p+1)+n1(G), n2(T)=n2(G), D(T)≥4 and Δ(T)=Δ(G). On the other hand, each k-(2)-vdc θ of G with k=χ″2s(G) corresponds to a total coloring φ of T by setting φ(x)=θ(x) for x∈V(G)∪(E(G)∖E′); φ(uu′)=φ(vv′)=θ(uv), φ(u′)=θ(v) and φ(v′)=θ(u) for uv∈E′. In general, this total coloring φ is not a (2)-vdc of T. Hence, max. This corollary follows from Lemmas 3–5 and Theorem 1.
Corollary 7. Let T be a spanning tree of a connected graph G , and G[E^*] be an induced graph over the edge subset E^* = E(G)\setminus E(T) . Then \chi\, ''_{2s}(G)\leq \chi\, ''_{2s}(T)+\chi\, '(G[E^*]) , where \chi\, '(H) is the chromatic index of a graph H .
A k -vdtc of a graph G is a total coloring f: V(G)\cup E(G)\rightarrow [1, k] such that C(w, f)\cup \{f(w)\}\neq C(z, f)\cup \{f(z)\} for distinct w, z\in V(G) , and \chi\, ''_s(G) is the least number k of colors required for which G has a k -vdtc. For 2\leq m\leq n , by Lemma 3, the difference \chi\, ''_{2s}(DS_{m+1, n+1})-\chi\, ''_{s}(DS_{m+1, n+1}) = m+n+1-(n+1) = m can be as large as we expect.
For a k -(2)-vdc f of a graph G , let S_i = \{f(x) = i:\; x\in V(G)\cup E(G)\} for i\in [1, k] . We call this coloring f to be an equitable k -(2)-vdc of G if two sizes of S_i and S_j differ at most one. Again, the smallest number k of colors required for which G has an equitable k -(2)-vdc is denoted as \chi \, ''_{2es}(G) . Clearly, \chi \, ''_{2s}(G)\leq \chi \, ''_{2es}(G) . We have shown that \chi \, ''_{2s}(H) = \chi \, ''_{2es}(H) for every 2 -star or every double star H . As further work we present a problem: Is there \chi \, ''_{2s}(T) = \chi \, ''_{2es}(T) for a tree T having D(T)\geq 4 and n_1(T)\neq \Delta(T) ? And we conjecture that: Let T be a tree with 2n_2(T)\leq (n_1(T)+k-1)^2 . Then \chi \, ''_{2s}(T)\leq n_1(T)+k .
Bing Yao was supported by the National Natural Science Foundation of China under Grant No. 61662066; Zhi-xiang Yin was supported by the National Natural Science Foundation of China under Grant Nos. 61672001 and 62072296.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
[1] |
P. N. Balister, A. Kostochka, H. Li, R. H. Schelp, Balanced edge colorings, J. Comb. Theory B, 90 (2004), 3–20. doi: 10.1016/S0095-8956(03)00073-X
![]() |
[2] |
P. N. Balister, B. Bollob\acute{a}s, R. H. Schelp, Vertex distinguishing coloring of graphs with \Delta(G) = 2, Discrete Math., 252 (2002), 17–29. doi: 10.1016/S0012-365X(01)00287-4
![]() |
[3] |
A. C. Burris, R. H. Schelp, Vertex-distingushing proper edge-colorings, J. Graph Theory, 26 (1997), 73–82. doi: 10.1002/(SICI)1097-0118(199710)26:2<73::AID-JGT2>3.0.CO;2-C
![]() |
[4] | J. Černý, M. Horňák, R. Soták, Observability of a graph, Math. Slovaca, 46 (1996), 21–31. |
[5] |
B. Yao, Z. F. Zhang, J. F. Wang, Some results on spanning trees, Acta Math. Appl. Sin. Engl. Ser., 26 (2010), 607–616. doi: 10.1007/s10255-010-0011-4
![]() |