Weakly edge-face coloring of Halin graphs

Menglei YU , Min CHEN

Front. Math. China ›› 2025, Vol. 20 ›› Issue (4) : 187 -197.

PDF
Front. Math. China ›› 2025, Vol. 20 ›› Issue (4) : 187 -197. DOI: 10.3868/s140-DDD-025-0015-x
RESEARCH ARTICLE

Weakly edge-face coloring of Halin graphs

Author information +
History +
PDF

Abstract

Let G=(V,E,F) be a connected loopless plane graph, with vertex set V, edge set E, and face set F. For any adjacent faces e1 and e2, if they are incident to the same face and appear consecutively on the edge of that face, then it is said that e1 and e2 are facially adjacent. A plane graph G is called weakly edge-face k-colorable indicating that there is a mapping π:EF{1,2,,k} such that any two incident edges and faces, adjacent faces, and facially adjacent edges receive distinct colors. The weakly edge-face chromatic number of G, denoted by χ¯ef(G), is defined to be the smallest integer k such that G has a weakly edge-face k-coloring. In 2016, Fabrici conjectured that every connected, loopless, and bridgeless plane graph was weakly edge-face 5-colorable. In this paper, a sufficient condition is provided for the foregoing conjecture to prove that Halin graphs are weakly edge-face 5-colorable in which the upper bound 5 is the best possible.

Graphical abstract

Keywords

Halin graph / wheel graph / weakly edge-face coloring / weakly edge-face chromatic number

Cite this article

Download citation ▾
Menglei YU, Min CHEN. Weakly edge-face coloring of Halin graphs. Front. Math. China, 2025, 20(4): 187-197 DOI:10.3868/s140-DDD-025-0015-x

登录浏览全文

4963

注册一个新账户 忘记密码

1 Introduction

All graphs considered in this paper are loopless connected graphs. Suppose G is a graph. If G can be drawn on a plane such that its edges intersect only at their vertices, then G is called a planarizable graph. Such drawing of the graph on a plane is called a planar embedding of the graph, or simply a plane graph. For a plane graph G, V(G), E(G), F(G), and Δ(G) are used to denote the vertex set, edge set, face set, and maximum degree of G, respectively. Halin graph G=TC is a plane graph, where T is a tree with a maximum degree of at least 3 and no vertices of degree 2, and C is a cycle formed by connecting the vertices of degree 1 in T in the order of the planar embedding. A wheel graph is a special type of Halin graph. Wn=K1+Cn1 is used to denote the wheel graph of order n.

The edge-face coloring problem of plane graphs was first studied independently by Jucovič [7] (1969) and Fiamčík [5] (1971). Let G be a plane graph. If k colors can be used to color the adjacent edges, adjacent faces, and incident edge and face of G with different colors, then G is said to be edge-face k-colorable. The edge-face chromatic number of G, denoted as χef(G), is defined as the minimum positive integer k such that G is edge-face k-colorable.

In 1975, Melnikov [8] proposed the conjecture that every plane graph G was edge-face (∆(G) + 3)-colorable. This conjecture was independently proven to be true by Waller [13], Sanders and Zhao [9], and Wang and Lih [14] successively. After Melnikov's conjecture was proven, Sanders and Zhao [11] proposed a stronger conjecture in 2001: every plane graph G was edge-face (Δ(G)+2)-colorable. It was verified that the conjecture held for Δ(G)3 and Δ(G)7 in references [10] and [11], respectively. In 2014, Chen et al. [2] completely solved the case of Δ(G)=6. Therefore, the cases of Δ(G)=4 and Δ(G)=5 remain challenging problems at present. Regarding Halin graphs, Gong and Wu [6] proved that if 3Δ(G)5, then χef(G)5; if Δ(G)6, then χef(G)=Δ(G).

In recent years, based on edge-face coloring [12], Fabrici et al. [4] proposed a new coloring concept, namely, the weakly edge-face coloring of plane graphs. For any two adjacent edges e1 and e2, if they are incident to the same face and appear consecutively on the edge of that face, then e1 and e2 are said to be face adjacent. A graph G is weakly edge-face k-colorable indicating that there exists a mapping π:EF{1,,k} such that any two incident edges and faces, adjacent faces, and face-adjacent edges receive distinct colors. The weak edge-face chromatic number of plane graph G, denoted as χ¯ef(G), is the minimum value of k for which G is weakly edge-face k-colorable. Obviously, χ¯ef(G)χef(G). When proposing the concept, Fabrici et al. [4] proved that connected, loopless, and bridgeless plane graphs were weakly edge-face 6-colorable and put forward the following conjecture.

Conjecture 1.1 [4]  Every connected, loopless, and bridgeless plane graph is weakly edge-face 5-colorable. In addition, they also gave the weak edge-face chromatic number of wheel graphs.

Theorem 1.1 [4]  For every wheel graph Wn, where n3, there is

χ¯ef(Wn)={5,n=2k+1,k1;4,n=2k,k2.

In this paper, we study the weak edge-face chromatic numbers of Halin graphs and obtain the following conclusions, which illustrates that Halin graphs satisfy Conjecture 1.1 and the upper bound 5 is optimal.

Theorem 1.2  Halin graphs are weakly edge-face 5-colorable.

2 Lemmas and notations

Let G be a Halin graph. Suppose vV(G). The number of edges incident to v is called the degree of v, denoted as dG(v). The set of all vertices adjacent to v is called the neighborhood of v in G, denoted as NG(v). For convenience, the unbounded face formed by V(C) is called the outer face, denoted as fo, and the remaining faces are called inner faces. The vertices in V(fo) are called outer vertices, and the vertices in GV(fo) are called inner vertices. Denote the set of all inner vertices as Vint(G), that is Vint(G)=V(G)V(fo). We call the inner vertices adjacent to outer vertices special inner vertices. Obviously, there are at least two special inner vertices in Halin graphs (except wheel graphs). For any edge eE(fo), denote the inner face incident to fo at the edge e as feo. Notations not defined in this paper are all cited from the book by Bondy and Murty [1].

In reference [3], Chen and Wang give a complete characterization of the structure of Halin graphs.

Lemma 2.1 [3]  LetG=TCbe any Halin graph other than a wheel graph. ThenGmust contain one of the following three structures, as shown inFig. 1.

(A1) There exists a special inner 3-vertex v, an inner vertex u, and a path xv1v2y on C such that NG(v)={v1,v2,u} and uyE(G);

(A2) There exist special inner 3-vertices v, w, an inner vertex u, and a path xv1v2w1w2y on C such that NG(v)={v1,v2,u} and NG(w)={w1,w2,u};

(A3) There exists a special inner (t+1)-vertex v, an inner vertex u, and a path xv1v2vty on C such that NG(v)={v1,v2,,vt,u}.

3 Proof of Theorem 1.2

Suppose G is a Halin graph. When Δ(G)5, then χ¯ef(G)χef(G)5. Therefore, in the following, it is assumed that Δ(G)6. Mathematical induction on |Vint(G)| is used to complete the proof of Theorem 1.2.

When |Vint(G)|=1, G is a wheel graph. By Theorem 1.1, χ¯ef(G)5, so Theorem 1.2 holds. Now it is assumed that when |Vint(G)|=k, where k ≥ 2, Theorem 1.2 holds. Next, consider the case when |Vint(G)|=k+1. Denote the color set C={1,2,,5}. According to Lemma 2.1, the following three cases are discussed.

Case 1: G contains the structure (A1).

Let G=G{v1,v2}+{xv,yv}. Obviously, G* is a Halin graph, satisfying Δ(G)=Δ(G), and |Vint(G)|=|Vint(G)|1=k. According to the induction hypothesis, G* has a weak edge-face 5-coloring π*. Denote the outer face of G* as *, and the face incident to f* at the edge e as fe. Next, extend π* to G so that G has a weak edge-face 5-coloring π. Without losing generality, assume π(f)=1. Denote A={xv1,fxv1o,yv2,fyv2o} and S={vv1,v1v2,vv2,fv1v2o}.

Firstly, let π(fo)=1. Secondly, assign the colors of xv,fxv,yv,fyv in G* to the corresponding elements xv1,fxv1o,yv2,fyv2o in G. Then, let π(s)=π(s), where sE(G)F(G)(AS{fo}). Next, it requires to show that the elements in S can also be properly colored so that G is weakly edge-face 5-colorable. For convenience, denote yNG(y){v2,u}. Let uu,uu be face adjacent to the edges uv, uy, respectively. It should be noted that when dG(u)=3, u and u coincide. π(A) is used to represent the set of colors of all elements in A at G. Obviously, 2|π(A)|4 and1π(A). We will discuss it in three sub-cases according to the value of |π(A)|.

Case 1.1: |π(A)|=4.

Without losing generality, assume that π(xv1)=2, π(fxv1o)=3, π(yv2)=4, and π(fyv2o)=5. This implies that π(uv)=1. It requires to color vv1 with color 5, v1v2 with color 3, vv2 with color 2, and fv1v2o with color 4. It is verified that the resulting coloring is a weakly edge-face 5-coloring of G.

Case 1.2: |π(A)|=3.

For the convenience, denote s1=vv1,s2=v1v2,s3=vv2, and s4=fv1v2o, that is S={s1,s2,s3,s4}. For any i{1,2,3,4}, let A(si) represent the set of available colors for the element si in G. Let A(S)=i=1A(si). Assertion 3.1 is first introduced.

Assertion 3.1  Assume |A(si)|=2(1i4) and |A(S)|=4. If for any {i,j,k}{1,2,3,4}, at most two of A(si), A(sj), A(sk) are the same, then there must exist a coloring π:SA(S) such that π(si)A(si)(1i4) and for any si, sjS satisfies π(si)π(sj).

Proof Firstly, assume that there exist A(si) and A(sj) such that A(si)=A(sj). By symmetry, without losing generality, let A(s1)=A(s2)={a,b}. According to the problem statement, A(s3){a,b} and A(s4){a,b}. This implies that there exist c1A(s3){a,b} and c2A(s4){a,b} such that c1c2; otherwise, it can be deduced that |A(S)|=3, which contradicts the given conditions. At this point, it requires to color s1, s2, s3, s4 with colors a, b, c1, c2, respectively.

Secondly, assume that there exist A(si) and A(sj) such that A(si)A(sj)=. Without losing generality, let A(s1)={a,b} and A(s2)={c,d}. Further, by symmetry, it is assumed A(s3)={a,c}. If cA(s4), then color s3 with c, s2 with d, s4 with αA(s4){c,d}, and s1 with βA(s1){α}. If cA(s4), then it can be deduced that A(s4)={b,c}, and then color s1, s2, s3, s4 with a, d, c, b, respectively.

Finally, it is assumed that for any {i,j}{1,2,3,4}, |A(si)A(sj)|=1. Since |A(S)|=4, it is not difficult to deduce that |A(s1)A(s2)A(s3)A(s4)|=0. Therefore, according to the principle of inclusion-exclusion

|A(S)|=i=14|A(si)|1i<j4|A(si)A(sj)|+1i<j<k4|A(si)A(sj)A(sk)||i=1A(si)|,

it is known

1i<j<k4|A(si)A(sj)A(sk)|=2.

This contradicts the fact 1i<j<k4|A(si)A(sj)A(sk)|1, thus completing the proof of Assertion 3.1. □

Next, according to the positional relationship, consider Case 1.2.1 and Case 1.2.2.

Case 1.2.1: π(xv1)=π(fyv2o).

Without losing generality, assume π(xv1)=π(fyv2o)=2, π(fxv1o)=3, and π(yv2)=4. It is found that the color of the edge uv is either 1 or 5. If π(uv)=5, color s1 with color 4, s2 with color 3, s3 with color 1, and s4 with color 5. Now consider the case when π(uv)=1.

Case 1.2.1a: π(fyyo)5 and π(uy)5.

Change the color of fyv2o to 5, and there are A(s1)={4,5}, A(s2)={3,5}, A(s3)={2,3}, and A(s4)={2,4}. It is verified that Ai(1i4) satisfies Assertion 2.1. Therefore, G has a weakly edge-face 5-coloring.

Case 1.2.1b: π(fyyo)=5 and π(uy)5.

It is found that π(uy)=3. First, change the color of uv to αC{1,2,3,π(uu)}. If α = 5, then A(s1)={1,4}, A(s2)={3,5}, A(s3)={1,3}, and A(s4)={4,5}. A weakly edge-face 5-coloring of G is obtained according to Assertion 3.1. If α = 4, further change the color of yv2 to 5. Then there are A(s1)={1,5}, A(s2)={3,4}, A(s3)={1,3}, and A(s4)={4,5}, thus getting a weakly edge-face 5-coloring of G.

Case 1.2.1c: π(fyyo)5 and π(uy)=5.

If π(uu)4, then first change the color of uv to 4, and the subsequent discussion process is similar to the latter subcase of Case 1.2.1b. Now assume π(uu)=4. If π(uu)1, then swap the colors of uv and uy, so that A(s1)={1,4}, A(s2)={3,5}, A(s3)={1,3}, and A(s4)={4,5}. According to Assertion 3.1, a weakly edge-face 5-coloring of G can be obtained. Now assume π(uu)=1. If it is possible to change the color of uy to 2 or 3, then further change the color of fyv2o to 5, that is, A(s1)={4,5}, A(s2)={3,5}, A(s3)={2,3}, and A(s4)={2,4}. Then, by applying Assertion 3.1, a weakly edge-face 5-coloring of G can be constructed. Otherwise, there must be that π(yy)=2 and π(fyyo)=3. At this point, change the colors of both yv2 and uv to 5 and then change the color of uy to 4. As a result, there are A(s1)={1,4}, A(s2)={3,4}, A(s3)={1,3,4}, and A(s4)={4,5}. Let A(s3)={1,3}. It is verified that A(s1), A(s2), A(s3), A(s4) satisfy Assertion 3.1. Q.E.D.

Case 1.2.2: π(yv2)=π(fxv1o).

Without losing generality, assume π(yv2)=π(fxv1o)=2, π(xv1)=3, and π(fyv2o)=4. At this point, π(uv){1,5}. If π(uv)=5, color vv1 with color 4, v1v2 with color 5, vv2 with color 1, and fv1v2o with color 3.

Next, assume π(uv)=1. For m{3,5}, if m{π(yy),π(uy)}, then change the color of yv2 to m. If m = 5, then A(s1)={4,5}, A(s2)={2,4}, A(s3)={2,3} and A(s4)={3,5}. If m = 3, then A(s1)={4,5}, A(s2)={2,4,5}, A(s3)={2,5}, and A(s4)={3,5}. It is verified that in these two cases, there is always Ai that satisfies the conditions of Assertion 3.1 and thus constructs a weakly edge-face 5-coloring of G. Now assume {π(yy),π(uy)}={3,5}. First, consider the case wherein π(yy)=3 and π(uy)=5. First change the color of uv to α{3,4}{π(uu)}. If α = 3, then A(s1)={1,4,5},A(s2)={4,5},A(s3)={1,5}, and A(s4)={3,5}. If α = 4, then further change the color of fyv2o to 3 and the color of yv2 to 4. At this time, A(s1)={1,5}, A(s2)={2,5},A(s3)={1,2,5}, and A(s4)={4,5}. In both cases, there is available color set that meets the conditions of Assertion 3.1 so that G has a weakly edge-face 5-coloring. Now consider the case wherein π(yy)=5 and π(uy)=3. Similarly, change the color of uv to α{4,5}{π(uu)}. If α = 5, there are A(s1)={1,4},A(s2)={4,5}, A(s3)={1,3}, and A(s4)={3,5}. If \alpha^'=4, then further change the color of fyv2o to 5, and then we can deduce that A(s1)={1,5}, A(s2)={4,5}, A(s3)={1,3}, and A(s4)={3,4}. According to Assertion 3.1, construct a weakly edge-face 5-coloring of G.

Case 1.3: |π(A)|=2.

Without losing generality, assume π(xv1)=π(fyv2o)=2 and π(yv2)=π(fxv1o)=3. It is found that π(uv){1,4,5}.

Case 1.3a: π(uv)=1.

It is found π(uy){4,5}. Without losing generality, assume π(uy)=4. Then π(fyyo){3,5}. If π(fyyo)=3, then first change the color of fyv2o to 5 and then go back to Case 1.2.2. If π(fyyo)=5, then π(yy)=2. Change the color of yv2 to 5 and then return to Case 1.2.1.

Case 1.3b: π(uv){4,5}.

Without losing generality, assume π(uv)=4. If π(uy)5 and π(fyyo)5, then first change the color of fyv2o to 5, then there are A(s1)={1,5}, A(s2)={4,5}, A(s3)={1,2}, and A(s4)={2,4}. According to Assertion 3.1, construct a weakly edge-face 5-coloring of G. Now assume π(fyyo)=5 or π(uy)=5. If π(fyyo)=5, then first change the color of yv2 to 5, and there are A(s1)={1,5}, A(s2)={3,4}, A(s3)={1,3}, and A(s4)={4,5} and thus obtain a weakly edge-face 5-coloring of G according to Assertion 3.1. Now assume that π(uy)=5. If π(yy)=4, then there must be π(fyyo)=3. First, change the color of fyv2o to 5, and then color uy with a color different from 3, 4, 5, π(uu). Consequently, there are A(s1)={1,5}, A(s2)={4,5}, A(s3)={1,2}, and A(s4)={2,4}. If π(yy)4, then the color of yy can only be 2. First, change the color of yv2 to 4 and then obtain A(s1)={1,5}, A(s2)={3,5}, A(s3)={1,3,5}, and A(s4)={4,5}. It is verified that in these two cases there is always Ai that satisfies the conditions of Assertion 3.1 and then construct a weakly edge-face 5-coloring of G.

Case 2: G contains the structure (A2).

Let G=G{v1,v2,w1,w2}+{vx,vw,yw}. G* is a Halin graph, Δ(G)=Δ(G) and |Vint(G)|=|Vint(G)|2=k1. By the induction hypothesis, G* has a weakly edge-face 5-coloring π*. Similarly, denote the outer face of G* as f* and the face incident to f* at the edge e as fe. Without losing generality, assume π(f)=1. Similarly, extend π* to G such that G has a weakly edge-face 5-coloring π. Denote B={xv1,fxv1o,yw2,fyw2o} and S={vv1,v1v2,vv2,fv1v2o,v2w1,fv2w1o,ww1,w1w2,ww2,fw1w2o}.

Firstly, let π(fo)=1. Secondly, assign the colors xv,fxv,yw,fyw in G* to the corresponding elements xv1,fxv1o,yw2,fyw2o in G. After that, let π(s)=π(s), where sE(G)F(G)(BS{fo}). Next, it requires to show that the elements in S can also be properly colored so that G is weakly edge-face 5-colorable. Denote M={uv,vw,uw,fvw} again and use π(M) to represent the set of colors of all elements in M at G*. Obviously, |π(M)|=4. The two cases will be discussed following according to the occurrence of color 1.

Case 2.1: 1π(M).

At this time, π(M)={2,3,4,5}. Without losing generality, assume π(uv)=2, π(vw)=3, π(uw)=4, and π(fvw)=5. Firstly, assign the color π(fxv1o) to v1v2 and the color π(fyw2o) to w1w2. Secondly, color both vv2 and ww1 with color 1, color fv1v2o with color 2, color fw1w2o with color 4, and color fv2w1o with color 5. Finally, color vv1 with αC{1,2,π(xv1),π(fxv1o)}, color ww2 with βC{1,4,π(yw2),π(fyw2o)}, and color v2w1 with γC{1,5,π(fxv1o),π(fyw2o)}.

Case 2.2: 1π(M).

By symmetry, without losing generality, assume π(uv)=1, π(vw)=2, π(uw)=3, and π(fvw)=4. For the convenience, denote π(xv1)=α, π(fxv1o)=β, π(yw2)=γ, and π(fyw2o)=η. It is noted that αβ and γη. According to the occurrence of color 5, there are the following two cases.

(i) 5{γ,η}. This implies that γ = 4 and η = 2. First, assume 5{α,β}. Then α{3,4} and β{2,3}. Color v1v2,ww2,fv2w1o with color 5, fv1v2o with color α, vv2 with color β, ww1 with color 1, w1w2 with color 2, and fw1w2o with color 4. Then, since |C|=5, there are always colors being assigned to vv1 and v2w1 respectively, such that G is weakly edge-face 5-colorable.

Now assume 5{α,β}. First, consider the case wherein α=5. Color ww1 with color 1, w1w2 with color 2, and fw1w2o with color 3. Color v1v2,fv2w1o with color 4 simultaneously, color v2w1,ww2,fv1v2o with color 5 simultaneously, and color vv2with color β. Finally, color vv1 with a color different from 1 ,4, 5, β.

Now consider the case wherein β = 5. It is noted that α{3,4}.

If α = 3, then color ww2 with color 1, vv2, ww1 with color 2, color v2w1,fv1v2o,fw1w2o with color 3, color vv1,fv2w1o with color 4, and color v1v2, w1w2 with color 5.

If α = 4, first denote uu as the edge adjacent to the face uw and incident to fw2yo. Then change the color of uw to aC{1,2,3,π(uu)}. Next, color ww2 with color 1, vv1, v2w1 with color 2, color v1v2,w1w2,fv2w1o with color 3, color fv1v2o with color 4, and color vv2 with color 5. Finally, color fw1w2o with color a and color ww1 with b{4,5}{a}.

(ii) 5{γ,η}. First, assume γ = 5. Then there must be η = 2. If 5{α,β}, then color ww1 with color 1, w1w2 with color 2, fw1w2o with color 3, v2w1, ww2 with color 4, and v1v2,fv2w1o with color 5. Then color fv1v2o with color α, and vv2 with color β. Finally, color vv1 with a color different from 1, 5, α, β. Now assume 5{α,β}.

(1) Suppose α = 5. Then β{2,3}. We color ww1 with color 1, w1w2 with color 2, fw1w2o with color 3, vv1, ww2, fv2w1o with color 4, and v2w1,fv1v2o with color 5. Then color v1v2 with color β. Finally, color vv2 with a color different from 1, 4, 5, β.

(2) Suppose β = 5. Then α{3,4}. If α = 3, similarly color ww1 with color 1, vv1, w1w2 with color 2, v2w1, fv1v2o,fw1w2o with color 3, v1v2,ww2,fv2w1o with color 4, and finally color vv2 with color 5. If α = 4, denote uu as the edge adjacent to face uw and incident to fw2yo. First, color uw with a color different from 1, 2, 3, π(uu). Then color ww1 with color 1, vv1, w1w2 with color 2, ww2,v1v2,fv2w1o with color 3, fv1v2o,v2w1 with color 4, and fw1w2o,vv2 with color 5.

Now assume η = 5. Then there must be γ = 4.

(1) If 5{α,β}, then first color v1v2 with color 5, fv1v2o with color α, and vv2 with color β. Then color vv1 with a color different from 1, 5, α, β. If there exists a color aC{1,3,5,α,β}, then color fv2w1o with color a, and v2w1 with color bC{1,5,a,β}. It is found that a{2,4} and b{2,3,4}. Thus, a weakly edge-face 5-coloring of G can be obtained by coloring ww with color 1, ww2with color 2, fw1w2o with color 3, and w1w2 with color 5. Now assume that such a color a does not exist. Then there is only one possibility: that is, α = 4 and β = 2. At this time, first color ww1 with color 1, w1w2 with color 2, ww2, fv2w1o simultaneously with color 3, vw1, fw1w2o simultaneously with color 4, and then change the color of uw to a color different from 1, 3, 5, π(uu), where uu is the edge adjacent to the face uw and incident to fyw2o.

(2) Now assume 5{α,β}. If β = 5, then first assign the color α to fv1v2o. Then color ww1 with color 1, vv1, ww2, fv2w1o with color 2, fw1w2o with color 4, and vv2,w1w2 with color 5. Finally, color v1v2 with aC{1,2,5,α}, and then color v2w1 with bC{1,2,5,a}. Next, consider the case wherein α = 5. If β = 3, color ww1 with color 1, vv1, ww2, fv2w1o with color 2, vv2, fw1w2o with color 3, v2w1, fv1v2o with color 4, and v1v2,w1w2 with color 5. If β = 2, color ww1 with color 1, vv2,ww2 with color 2, vv1 ,v2w1, fw1w2o with color 3, v1v2, fv2w1o with color 4, and w1w2, fv1v2o with color 5.

Case 3: G contains the structure (A3).

Let G=G{v1,v2,,vt}+{vx,vy}. G* is a Halin graph, Δ(G)=Δ(G) and |Vint(G)|<|Vint(G)|. By the induction hypothesis, G* has a weakly edge-face 5-coloring π*. Similarly, denote the outer face of G* as f*, and the face incident to f* at the edge e as fe. Next, extend π* to G such that G has a weakly edge-face 5-coloring π. Without losing generality, assume π(f)=1. Denote D={xv1,fxv1o,yvt,fyvto} and {vv1,vv2,,vvt,v1v2,v2v3,,vt1vt,fv1v2o,fv2v3o,,fvt1vto}.

Firstly, let π(fo)=1. Secondly, assign the colors of xv,fxv,fxv,yv,fyv in G* to the corresponding elements xv1,fxv1o,yvt,fyvto in G. Then, let π(s)=π(s), where sE(G)F(G)(DS{fo}). It requires to prove that the elements in S can also be colored properly so that G is weakly edge-face 5-colorable. Denote π(D) as the color set for all elements in D at G. It is found that 2|π(D)|4 and 1π(D). The following will discuss the three cases.

Case 3.1: |π(D)|=4.

Without losing generality, assume π(xv1)=2, π(fxv1o)=3, π(yvt)=4, and π(fyvto)=5. This indicates that π(vu)=1. If t = 3, color vv3, fv1v2o with color 2, vv2 with color 3, v1v2, fv2v3o with color 4, and color vv1, v2v3 with color 5. It is verified that the resulting coloring is a weakly edge-face 5-coloring of G. Now assume t ≥ 4. First, color vvt, fv1v2o with color 2, v1v2 with color 3, vv1, fvt1vto with color 4, and vt1vt, fvt2vt1o with color 5. Then, when t is odd, color vvt1 with color 1; otherwise, color vvt1 with color 3. Next, color the uncolored edges and faces in S to obtain a weakly edge-face 5-coloring of G. For the edges, let 2it2. If i is even, color the edges vvi and vivi+1 with colors 1 and 4 respectively. If i is odd, color the edges vvi and vivi+1 with colors 2 and 3 respectively. For the faces, let 2jt3. If j is even, color the face fvjvj+1o with color 3. If j is odd, color the face fvjvj+1o with color 4.

Case 3.2: |π(D)|=3.

Without losing generality, assume π(xv1)=π(fyvto)=2, π(fxv1o)=3, and π(yvt)=4. It is found that π(uv){1,5}. Firstly, color fvt1vto with color 4. Secondly, color vvt1 with the same color as uv, and color vvt with a{1,5}{π(uv)}. Next, let 1it1. If i is even, color the edge vivi+1 with color 2. If i is odd, color the edge vivi+1 with color 3. Then, let 1jt2. If j is odd, color the edge vvj with color a and the face fvjvj+1o with color 2. If j is even, color the edge vvj with color 4 and the face fvjvj+1o with color 3.

Case 3.3: |π(D)|=2.

Without losing generality, assume π(xv1)=π(fyυto)=2 and π(yvt)=π(fxυ1o)=3. It is known that π(uv){1,4,5}. Now, extend π to G in different cases according to the value of t to complete the proof of Theorem 1.2.

(1) t = 3. First, color fv2v3o with color 3. Then color v2v3 and fv1v2o with color 4. After that, color vv1 and vv3 with b{1,5}{π(uv)}. Finally, color v1v2 with color 3 and vv2 with color 2.

(2) t ≥ 4. First, color fvt1vto with color 3 and vvt1 with π(uv). Then color vt1vt, fvt2vt1 with a{4,5}{π(uv)}. Next, color vvt2, vvt with b{1,4,5}{a,π(uv)}. Now, let 1it2. If i is odd, color the edge vivi+1 with color 3. If i is even, color the edge vivi+1 with color 2. Then, let 1jt3. If j is odd, color the edge vvj with color a and the face fvjvj+1o with color 2. If j is even, color the edge vvj with color π(uv) and the face fvjvj+1o with color 3.

References

[1]

BondyJ.A. , Murty, U.S.R., Graph Theory, London: Springer-Verlag, 2008

[2]

Chen M., Raspaud, A. , Wang, W.F. . Plane graphs with maximum degree 6 are edge-face 8-colorable. Graphs Combin. 2014; 30: 861–874

[3]

Chen M. , Wang, W.F. . The 2-dipath chromatic number of Halin graphs. Inform. Process. Lett. 2006; 99: 47–53

[4]

Fabrici I., Jendrol, S. , Vrbjarová, M. . Facial entire colouring of plane graphs. Discrete Math. 2016; 339: 626–631

[5]

Fiamčík . Simultaneous coloring of 4-valent maps. Math. Slovaca 1971; 21: 9–13

[6]

Gong Q.Y. , Wu, J.L. . Edge-face coloring of Halin graphs. Inform. Tech. 2008; 7: 64–67

[7]

Jucovič . On a problem in map coloring. Math. Slovaca 1969; 19: 225–227

[8]

MelnikovL.S., Problem 9, In: Recent Advances in Graph Theory (Fiedler, M. ed.), Proceedings of the Second Czechoslovak Symposium, Prague, June, 1974, Academia Prague, Czechoslovak, 1975, p. 543

[9]

Sanders D.P. , Zhao, Y. . On simultaneous edge-face colorings of plane graphs. Combinatorica 1997; 17: 441–445

[10]

Sanders D.P. , Zhao, Y. . A five-color theorem. Discrete Math. 2000; 220: 279–281

[11]

Sanders D.P. , Zhao, Y. . On improving the edge-face coloring theorem. Graphs Combin. 2001; 17: 329341

[12]

Sereni J. , Stehlik, M. . Edge-face colouring of plane graphs with maximum degree nine. J. Graph Theory 2011; 66: 332–346

[13]

Waller A.O. . Simultaneous colouring the edges and faces of plane graphs. J. Combin. Theory Ser. B 1997; 69: 219–221

[14]

Wang W.F. , Lih, K. . A new proof of Melnikov’s conjecture on the edge-face coloring of plane graphs. Discrete Math. 2002; 253: 87–95

RIGHTS & PERMISSIONS

Higher Education Press 2025

AI Summary AI Mindmap
PDF

14

Accesses

0

Citation

Detail

Sections
Recommended

AI思维导图

/