
An edge-coloring of a graph G with colors 1,…,t is an interval t-coloring if all colors are used and the colors of edges incident to each vertex of G are distinct and form an interval of integers. A graph G is interval colorable if it has an interval t-coloring for some positive integer t. For an interval colorable graph G, the least and the greatest values of t for which G has an interval t-coloring are denoted by w(G) and W(G). Let G be a graph with vertex set V(G)={u1,…,um}, m≥2, and let hm=(Hi)i∈{1,…,m} be a sequence of vertex-disjoint with V(Hi)={x(i)1,…,x(i)ni}, ni≥1. The generalized lexicographic products G[hm] of G and hm is a simple graph with vertex set ∪mi=1V(Hi), in which x(i)p is adjacent to x(j)q if and only if either ui=uj and x(i)px(i)q∈E(Hi) or uiuj∈E(G). In this paper, we obtain several sufficient conditions for the generalized lexicographic product G[hm] to have interval colorings. We also present some sharp bounds on w(G[hm]) and W(G[hm]) of G[hm].
Citation: Meiqin Jin, Ping Chen, Shuangliang Tian. Interval edge colorings of the generalized lexicographic product of some graphs[J]. AIMS Mathematics, 2024, 9(11): 30597-30611. doi: 10.3934/math.20241477
[1] | Shuangliang Tian, Ping Chen . Edge-coloring of generalized lexicographic product of graphs. AIMS Mathematics, 2024, 9(6): 15988-15995. doi: 10.3934/math.2024774 |
[2] | S. Gajavalli, A. Berin Greeni . On strong geodeticity in the lexicographic product of graphs. AIMS Mathematics, 2024, 9(8): 20367-20389. doi: 10.3934/math.2024991 |
[3] | Jin Cai, Shuangliang Tian, Lizhen Peng . On star and acyclic coloring of generalized lexicographic product of graphs. AIMS Mathematics, 2022, 7(8): 14270-14281. doi: 10.3934/math.2022786 |
[4] | Yanyi Li, Lily Chen . Injective edge coloring of generalized Petersen graphs. AIMS Mathematics, 2021, 6(8): 7929-7943. doi: 10.3934/math.2021460 |
[5] | Yunfeng Tang, Huixin Yin, Miaomiao Han . Star edge coloring of $ K_{2, t} $-free planar graphs. AIMS Mathematics, 2023, 8(6): 13154-13161. doi: 10.3934/math.2023664 |
[6] | Baolin Ma, Chao Yang . Distinguishing colorings of graphs and their subgraphs. AIMS Mathematics, 2023, 8(11): 26561-26573. doi: 10.3934/math.20231357 |
[7] | Ningge Huang, Lily Chen . AVD edge-colorings of cubic Halin graphs. AIMS Mathematics, 2023, 8(11): 27820-27839. doi: 10.3934/math.20231423 |
[8] | Yuehua Bu, Qi Jia, Hongguo Zhu, Junlei Zhu . Acyclic edge coloring of planar graphs. AIMS Mathematics, 2022, 7(6): 10828-10841. doi: 10.3934/math.2022605 |
[9] | Sakander Hayat, Hafiz Muhammad Afzal Siddiqui, Muhammad Imran, Hafiz Muhammad Ikhlaq, Jinde Cao . On the zero forcing number and propagation time of oriented graphs. AIMS Mathematics, 2021, 6(2): 1833-1850. doi: 10.3934/math.2021111 |
[10] | Yixin Zhang, Yanbo Zhang, Hexuan Zhi . A proof of a conjecture on matching-path connected size Ramsey number. AIMS Mathematics, 2023, 8(4): 8027-8033. doi: 10.3934/math.2023406 |
An edge-coloring of a graph G with colors 1,…,t is an interval t-coloring if all colors are used and the colors of edges incident to each vertex of G are distinct and form an interval of integers. A graph G is interval colorable if it has an interval t-coloring for some positive integer t. For an interval colorable graph G, the least and the greatest values of t for which G has an interval t-coloring are denoted by w(G) and W(G). Let G be a graph with vertex set V(G)={u1,…,um}, m≥2, and let hm=(Hi)i∈{1,…,m} be a sequence of vertex-disjoint with V(Hi)={x(i)1,…,x(i)ni}, ni≥1. The generalized lexicographic products G[hm] of G and hm is a simple graph with vertex set ∪mi=1V(Hi), in which x(i)p is adjacent to x(j)q if and only if either ui=uj and x(i)px(i)q∈E(Hi) or uiuj∈E(G). In this paper, we obtain several sufficient conditions for the generalized lexicographic product G[hm] to have interval colorings. We also present some sharp bounds on w(G[hm]) and W(G[hm]) of G[hm].
All graphs considered in this paper are finite, undirected, and simple graphs. Let V(G) and E(G) denote the sets of vertices and edges of graph G, respectively. The degree of a vertex v in G is denoted by dG(v), the maximum degree of G by Δ(G), and the complement of the graph G by ¯G. We denote by [a,b] the interval of integers {a,…,b}, and by (x)k the x(modk), where x is an integer and k is a positive integer. The terms and concepts that we do not define can be found in [3].
The notion of interval colorings was introduced in 1987 by Asratian and Kamalian [1]. This concept was introduced for studying the problems that are related to constructing timetables without "gaps" for teachers and classes. Hansen [6] presented a similar scenario: a school can schedule parent-teacher conferences in consecutive time slots if the bipartite graph, with parents and teachers as vertices and meetings as edges, has an interval coloring.
A proper edge coloring of a graph G is a coloring of the edges of G such that no two adjacent edges receive the same color. The chromatic index χ′(G) of a graph G is the minimum number of colors used in a proper edge coloring of G. If α is a proper edge coloring of G and v∈V(G), we denote by S(v,α) the set of colors of edges incident to v.
A proper edge coloring of a graph G with colors 1,…,t is an interval t-coloring if all colors are used, and for any vertex v of G, the set S(v,α) is an interval of consecutive integers. A graph G is interval colorable if it has an interval t-coloring for some positive integer t. The set of interval colorable graphs is denoted by N. For a graph G∈N, the least and the greatest values of t for which G has an interval t-coloring are denoted by w(G) and W(G), respectively.
The Vizing Theorem[19] states that χ′(G)=Δ(G) or χ′(G)=Δ(G)+1. Graphs with χ′(G)=Δ(G) are said to be Class 1; graphs with χ′(G)=Δ(G)+1 are said to be Class 2. Asratian and Kamalian [1] proved that if G∈N, then χ′(G)=Δ(G), that is, G is said to be Class 1. Not all Class 1 graphs are interval colorable; even a simple graph such as wheel Wn(n≥3) is not necessarily interval edge colorable. Giaro et al.[4] proved that Wn∈N if and only if n=4,7,10; otherwise, Wn∉N. Asratian et al.[2], Holyer [7], and Sevastjanov [16] proved that it is a NP-complete problem to decide whether a regular graph or a bipartite graph has an interval coloring or not. For a graph G∈N, the exact values of the parameters W(G) and w(G) are known only for paths, even cycles, trees, complete bipartite graphs [1,9], and Möbius ladders [12]. However, for some common interval-colorable graphs, such as complete graphs and n-dimensional cubes [13], the exact upper and lower bounds on the number of colors remain unknown.
We know that the generalized lexicographic product of graphs is one of the important tools for constructing classes of graphs in graph theory. Common graph operations, such as the join of graphs and the lexicographic product of graphs, are essentially special cases of the generalized lexicographic product. However, to the best knowledge of the author, there are no studies related to the interval edge coloring of generalized lexicographic products. This fact motivates us to begin an exploration of the interval edge coloring of the generalized lexicographic product of graphs.
The generalized lexicographic product of graphs is defined as follows [17]: let G be a graph with vertex set V(G)={u1,…,um}, m≥2, and let hm=(Hi)i∈{1,…,m} be a sequence of vertex-disjoint with V(Hi)={x(i)1,…,x(i)ni}, ni≥1. The generalized lexicographic products G[hm] of G and hm is a simple graph with vertex set ∪mi=1V(Hi), in which x(i)p is adjacent to x(j)q if and only if either ui=uj and x(i)px(i)q∈E(Hi) or uiuj∈E(G). If Hi≅H for every i∈{1,…,m}, then G[hm] becomes the lexicographic product of G and H, denoted by G[H]. If G≅K2, then G[h2] denotes a join H1+H2 of vertex disjoint graphs H1 and H2. If G≅Km, and every graph Hi is the complement of a complete graph with ni vertices (that is, Hi≅¯kni), then G[hm] denotes a complete m-partite graph Kn1,…,nm. For convenience, we write ˜G to denote G[hm].
Let G(1)=G[¯Kn] and G(2)=⋃mi=1Hi, then ˜G can be decomposed into the union of two edge-disjoint subgraphs G(1) and G(2), that is,
˜G=G(1)∪G(2). |
Figure 1 shows the decomposition of ˜G=W7[h7], where h7=(Hi)i∈{1,…,7} is a sequence of vertex-disjoint graphs, each with n vertices.
Kamalian [10] proved that the complete bipartite graph Km,n has an interval t-coloring if and only if m+n−gcd(m,n)≤t≤m+n−1. Grzesik and Khachatrian [5] proved that complete tripartite graphs K1,m,n are interval colorable if and only if gcd(m+1,n+1)=1. Puning Jing et al.[8] extended the result and obtained several sufficient conditions for a complete tripartite graph Kl,m,n to admit an interval coloring, where gcd(m,n) is the greatest common divisor of m and n. Petrosyan [14] investigated interval edge colorings of lexicographic products and obtained the following two results.
Theorem 1.1. If G∈N, then G[¯kn]∈N, and w(G[¯kn])≤w(G)n, W(G[¯kn])≥(W(G)+1)n−1.
Theorem 1.2. If G,H∈N and H is a r-regular graph, then G[H]∈N, and w(G[H])≤w(G)|V(H)|+r, W(G[H])≥W(G)|V(H)|+r.
Yepremyan et al. [15] proved that if G is a tree and H is a path or a star, then G[H]∈N. Tepanyan et al. [18] proved that if G∈N, and H is a regular graph, complete bipartite graph, or tree, then G[H]∈N. They obtained the following results:
Theorem 1.3. For all positive integers n≥2, if G∈N, then G[C2n]∈N, and w(G[C2n])≤2(w(G)n+1), W(G[C2n])≥(2W(G)+1)n+1.
In this paper, our goal is to find sufficient conditions for a generalized lexicographic product ˜G to admit an interval coloring. Moreover, we also hope to present some sharp bounds on w(˜G) and W(˜G), where G is an interval colorable graph with m vertices and all graphs in hm have n vertices, n≥4.
The following two lemmas will be used later:
Lemma 1.1. [11] If α is an edge-coloring of a connected graph G with colors 1,…,t such that the edges incident to each vertex v∈V(G) are colored by distinct and consecutive colors and min{α(uiuj)|uiuj∈E(G)}=1, max{α(uiuj)|uiuj∈E(G)}=t, then α is an interval t-coloring of G.
Lemma 1.2. Let α be an interval t-coloring of G, and let β(uiuj)=t−α(uiuj)+1 for every edge uiuj∈E(G), then β is also an interval t-coloring of G.
Proof. For any vertex ui∈V(G), we assume that S(ui,α)=[a,a+d(ui)−1]. By the definition of β, we have S(ui,β)=[t−a−d(ui)+2,t−a+1], that is, β is an interval t-coloring of G.
In Section 2, we obtained several sufficient conditions for G[hm] to admit an interval edge coloring, and we shall present the bounds of w(˜G) and W(˜G).
Let G∈N. If there exists a graph Hi in hm that is not isomorphic to an empty graph, then let w(Hk0)=max{w(Hr)|Hr∈N,r=1,…,m}. If Hi is isomorphic to a path, then let Hi=x(i)1…x(i)n; if Hi is isomorphic to a cycle, then let Hi=x(i)1…x(i)nx(i)1.
We define the following three classes of graphs:
(i) G1: each graph Hi in hm is isomorphic to a path or an empty graph;
(ii) G2: each graph Hi in hm is isomorphic to an empty graph or an interval-colorable regular graph, but not all graphs in hm are empty graphs;
(iii) G3: each graph Hi in hm is isomorphic to a path or a cycle or an empty graph of even order, and Hk0 is a cycle.
Let G=G1∪G2∪G3. And let ˜G∈Gz, where z=1,2,3. If there exists an interval t-coloring α satisfying the condition I, then we say that ˜G belongs to the subclass G1z of Gz, otherwise, we say that ˜G belongs to the subclass G2z of Gz. Clearly, we have
Gz=G1z∪G2z,z=1,2,3. |
Condition I. There exists an edge ui0uj0∈E(G) with α(ui0uj0)=1 or α(ui0uj0)=t, such that Hi0≅Hk0 or Hj0≅Hk0.
Since G∈N, there exists an interval w(G)-coloring α1 and an interval W(G)-coloring α2 of G. For any vertex ui∈V(G), denote by τi(αl) the minS(ui,αl), and by τ′i(αl) the maxS(ui,αl), where l=1,2, i=1,…,m.
We extend Theorem 1.1 from the lexicographic product of graphs to the generalized lexicographic product of graphs and obtain the following result:
Theorem 2.1. If ˜G∈G1, then ˜G∈N. Moreover,
w(˜G)≤w(G)n+2(λ)2+(λ+1)2(n)2,W(˜G)≥(W(G)+1)n−λ, |
where, λ=1 if ˜G∈G11, λ=2 if ˜G∈G21.
Proof. If Hi is isomorphic to a path, then let Hi=x(i)1…x(i)n. For the proof, we consider the following two cases:
Case 1. ˜G∈G11.
Then Condition I holds and λ=1. By Lemma 1.2, we can assume that α(ui0uj0)=1. Obviously, Hk0≅Pn, w(G)n+2(λ)2+(λ+1)2(n)2=w(G)n+2,(W(G)+1)n−λ=(W(G)+1)n−1.
Now, we prove that w(˜G)≤w(G)n+2. For this, we define an edge-coloring β1 of the graph ˜G in the following two steps:
Step 1. For every edge x(i)px(j)q∈E(G(1)), 1≤p,q≤n, if (n)2=1, then let
β1(x(i)px(j)q)={(α1(uiuj)−1)n+q+1,p=1;(α1(uiuj)−1)n+(p+q−3)n+3,2≤p≤n. |
Otherwise, let
β1(x(i)px(j)q)={(α1(uiuj)−1)n+(p+q−1)n+2,p+q≠n+1;α1(uiuj)n+2,p+q=n+1. |
Step 2. For every edge x(i)px(i)p+1∈E(G(2)), 1≤p≤n−1, let
β1(x(i)px(i)p+1)={(τi(α1)−1)n+(p+1)2+1,(n)2=1;(τi(α1)−1)n+(p)2+1,(n)2=0. |
Let us prove that β1 is an interval edge coloring of the graph ˜G.
First, we prove that the set S(x(i)p,β1) is an interval for each vertex x(i)p∈V(˜G), where 1≤i≤m, 1≤p≤n.
If (n)2=1 and Hi is isomorphic to a path, by the definition of β1, we have
S(x(i)1,β1)={(τi(α1)−1)n+1+1,…,(τ′i(α1)−1)n+n+1}∪{(τi(α1)−1)n+1}=[(τi(α1)−1)n+1,τ′i(α1)n+1]; |
S(x(i)p,β1)={(τi(α1)−1)n+3,…,(τ′i(α1)−1)n+(n−1)+3}∪{(τi(α1)−1)n+(p+1)2+1}∪{(τi(α1)−1)n+(p)2+1=[(τi(α1)−1)n+1,τ′i(α1)n+2],2≤p≤n−1; |
S(x(i)n,β1)={(τi(α1)−1)n+3,…,(τ′i(α1)−1)n+(n−1)+3}∪{(τi(α1)−1)n+(n+1)2+1}=[(τi(α1)−1)n+2,τ′i(α1)n+2]; |
If (n)2=1 and Hi≅¯Kn, by the definition of β1, we have
S(x(i)1,β1)={(τi(α1)−1)n+1+1,…,(τ′i(α1)−1)n+n+1}=[(τi(α1)−1)n+2,τ′i(α1)n+1]; |
S(x(i)p,β1)={(τi(α1)−1)n+3,…,(τ′i(α1)−1)n+(n−1)+3}=[(τi(α1)−1)n+3,τ′i(α1)n+2],2≤p≤n. |
If (n)2=0 and Hi is isomorphic to a path, by the definition of β1, we have
S(x(i)1,β1)=S(x(i)n,β1)={(τi(α1)−1)n+1+1,…,τ′i(α1)n+2}∪{(τi(α1)−1)n+2}=[(τi(α1)−1)n+2,τ′i(α1)n+2]; |
S(x(i)p,β1)={(τi(α1)−1)n+2,…,τ′i(α1)n+2}∪{(τi(α1)−1)n+(p−1)2+1}∪{(τi(α1)−1)n+(p)2+1=[(τi(α1)−1)n+1,τ′i(α1)n+2],2≤p≤n−1. |
If (n)2=0 and Hi≅¯Kn, by the definition of β1, we have
S(x(i)1,β1)=S(x(i)n,β1)={(τi(α1)−1)n+1+2,…,τ′i(α1)n+2}=[(τi(α1)−1)n+3,τ′i(α1)n+2]; |
S(x(i)p,β1)={(τi(α1)−1)n+1+2,…,τ′i(α1)n+2}=[(τi(α1)−1)n+3,τ′i(α1)n+2],2≤p≤n−1. |
Second, note that
d˜G(x(i)p)=dH(x(i)p)+dG(x(i)p)n={(τ′i(α1)−τi(α1))n+n,Hi≅¯Kn;(τ′i(α1)−τi(α1))n+n+1,Hi≅Pn,andp=1orp=n;(τ′i(α1)−τi(α1))n+n+2,Hi≅Pn,and2≤p≤n−1. |
Clearly, we have
maxS(x(i)p,β1)−minS(x(i)p,β1)=d˜G(x(i)p)−1,1≤p≤n. |
This implies that β1 is a proper edge coloring of ˜G.
Finally, we show that in the coloring β1, all colors are used. Clearly, there exists an edge x(i0)p0x(i0)p0+1∈E(˜G) such that β1(x(i0)p0x(i0)p0+1)=1, since in the coloring α1 there exists a vertex ui0 such that τi0(α1)=1, and since β1(x(i0)1x(i0)2)=(τi0(α1)−1)n+1 when (n)2=1 and β1(x(i0)2x(i0)3)=(τi0(α1)−1)n+1 when (n)2=0. Similarly, there exists an edge x(i1)p1x(j1)q1∈E(˜G) such that β1(x(i1)p1x(j1)q1)=w(G)n+2, since in the coloring α1 there exists an edge ui1uj1∈E(G) with α1(ui1uj1)=w(G), and since β1(x(i1)2x(j1)n)=(α1(ui1uj1)−1)n+(n+2−3)n+3=w(G)n+2 when (n)2=1 and β1(x(i1)1x(j1)n)=(α1(ui1uj1)−1)n+(n+1−3)n+3=w(G)n+2 when (n)2=0.
Now, by Lemma 1.1, we have that β1 is an interval (w(G)n+2)-edge coloring of the graph ˜G.
Next, we prove that W(˜G)≥(W(G)+1)n−1. For this, we define an edge-coloring β2 of the graph ˜G in the following two steps:
Step 1. For every edge x(i)px(j)q∈E(G(1)), 1≤p,q≤n, let
β2(x(i)px(j)q)={(α2(uiuj)−1)n+p+q,p+q≠2n;α2(uiuj)n,p+q=2n. |
Step 2. For every edge x(i)px(i)p+1∈E(G(2)), 1≤p≤n−1, let
β2(x(i)px(i)p+1)=(τi(α2)−1)n+p,p=1,⋯,n−1. |
Let us prove that β2 is an interval edge coloring of ˜G.
First, we prove that the set S(x(i)p,β2) is an interval for each vertex x(i)p∈V(˜G), where i=1,…,m, p=1,…,n.
If Hi is isomorphic to a path, by the definition of β2, we have
S(x(i)1,β2)={(τi(α2)−1)n+1+1,…,(τ′i(α2)−1)n+n+1}∪{(τi(α2)−1)n+1}=[(τi(α2)−1)n+1,τ′i(α2)n+1]; |
S(x(i)p,β2)={(τi(α2)−1)n+p+1,…,(τ′i(α2)−1)n+p+n}∪{(τi(α2)−1)n+p−1}∪{(τi(α2)−1)n+p=[(τi(α2)−1)n+p−1,τ′i(α2)n+p],2≤p≤n−1; |
S(x(i)n,β2)={τi(α2)n,…,(τ′i(α2)−1)n+n+n−1}∪{(τi(α2)−1)n+n−1}=[τi(α2)n,(τ′i(α2)+1)n−1]; |
If Hi≅¯Kn, by the definition of β2, we have
S(x(i)1,β2)={(τi−1)n+1+1,…,(τ′i−1)n+n+1}=[(τi−1)n+2,τ′in+1]; |
S(x(i)p,β2)={(τi(α2)−1)n+p+1,…,(τ′i(α2)−1)n+p+n}=[(τi(α2)−1)n+p+1,τ′i(α2)n+p],2≤p≤n−1; |
S(x(i)n,β2)={τi(α2)n,…,(τ′i(α2)−1)n+n+n−1}=[τi(α2)n,(τ′i(α2)+1)n−1]; |
Second, note that
d˜G(x(i)p)=dH(x(i)p)+dG(x(i)p)n={(τ′i(α2)−τi(α2))n+n,Hi≅¯Kn;(τ′i(α2)−τi(α2))n+n+1,Hi≅Pn,andp=1orp=n;(τ′i(α2)−τi(α2))n+n+2,Hi≅Pn,and2≤p≤n−1. |
Clearly, we have
maxS(x(i)p,β2)−minS(x(i)p,β2)=d˜G(x(i)p)−1,1≤p≤n. |
This implies that β2 is a proper edge coloring of ˜G.
Finally, we show that in the coloring β2, all colors are used. Clearly, there exists an edge x(i0)p0x(i0)p0+1∈E(˜G) such that β2(x(i0)p0x(i0)p0+1)=1, since in the coloring α2 there exists a vertex ui0 such that τi0(α2)=1, and since β2(x(i0)1x(i0)2)=(τi0(α2)−1)n+1. Similarly, there exists an edge x(i1)p1x(j1)q1∈E(˜G) such that β2(x(i1)p1x(j1)q1)=(W(G)+1)n−1, since in the coloring α2 there exists an edge ui1uj1∈E(G) with α2(ui1uj1)=W(G), and since β2(x(i1)n−1x(j1)n)=(α2(ui1uj1)−1)n+2n−1=(W(G)+1)n−1.
Now, by Lemma 1.1, we have that β2 is an interval ((W(G)n+1)n−1)-edge coloring of the graph ˜G.
Case 2. ˜G∈G21.
Then Condition I holds and λ=2. By Lemma 1.2, we can assume that α(ui0uj0)=1. It is easy to see that w(G)n+2(λ)2+(λ+1)2(n)2=w(G)n+(n)2,(W(G)+1)n−λ=(W(G)+1)n−2.
Now, we prove that w(˜G)≤w(G)n+(n)2. For this, we define an edge-coloring β3 of the graph ˜G in the following two steps:
Step 1. For every edge x(i)px(j)q∈E(G(1)), 1≤p,q≤n, if (n)2=1, then let
β3(x(i)px(j)q)={(α1(uiuj)−1)n+q,p=1;(α1(uiuj)−1)n+(p+q−3)n+2,2≤p≤n, |
if (n)2=0, then let
β3(x(i)px(j)q)={(α1(uiuj)−1)n+(p+q−1)n,p+q≠n+1;α1(uiuj)n,p+q=n+1. |
Step 2. For every edge x(i)px(i)p+1∈E(G(2)), 1≤p≤n−1, let
β3(x(i)px(i)p+1)={(τi(α1)−1)n+(p+1)2,(n)2=1;(τi(α1)−1)n−(p+1)2,(n)2=0. |
Let us prove that β3 is an interval (w(G)n+(n)2)-edge coloring of the graph ˜G.
First, for any edge x(i)px(j)q∈E(˜G), by the definitions of colorings β1 and β3, if (n)2=1, then β3(x(i)px(j)q)=β1(x(i)px(j)q)−1, otherwise, β3(x(i)px(j)q)=β1(x(i)px(j)q)−2. Since β1 is an interval edge coloring of ˜G, the coloring β3 is a proper edge coloring of ˜G and the color set of each vertex forms an integer interval.
Next, we show that in the coloring β3, all colors are used. Clearly, there exists an edge x(i0)p0x(j0)q0∈E(˜G) such that β3(x(i0)p0x(j0)q0)=1, since in the coloring α1 there exists a vertex ui0 such that τi0(α1)=1, and since β3(x(i0)1x(j0)1)=(τi0(α1)−1)n+1. Similarly, there exists an edge x(i1)p1x(j1)q1∈E(˜G) such that β3(x(i1)p1x(j1)q1)=w(G)n+(n)2, since in the coloring α1 there exists an edge ui1uj1∈E(G) with α1(ui1uj1)=w(G), and since β3(x(i1)2x(j1)n)=(w(G)−1)n+n+1=w(G)n+1 when (n)2=1 and β3(x(i1)1x(j1)n)=α1(ui1uj1)n=w(G)n when (n)2=0.
Now, by Lemma 1.1, we have that β3 is an interval (w(G)n+(n)2)-edge coloring of the graph ˜G.
Next, we prove that W(˜G)≥(W(G)+1)n−2. For this, we define an edge-coloring β4 of the graph ˜G in the following two steps:
Step 1. For every edge x(i)px(j)q∈E(G(1)), 1≤p,q≤n, let
β4(x(i)px(j)q)={(α2(uiuj)−1)n+p+q−1,p+q≠2n;α2(uiuj)n−1,p+q=2n. |
Step 2. For every edge x(i)px(i)p+1∈E(G(2)), 1≤p≤n−1, let
β4(x(i)px(i)p+1)=(τi(α2)−1)n+p−1. |
Let us prove that β4 is an interval ((W(G)+1)n−2)-edge coloring of the graph ˜G.
First, for any edge x(i)px(j)q∈E(˜G), by the definitions of colorings β2 and β4, we have β4(x(i)px(j)q)=β2(x(i)px(j)q)−1. Since β2 is an interval edge coloring of ˜G, the coloring β4 is a proper edge coloring of ˜G and the color set of each vertex forms an integer interval.
Next, we show that in the coloring β4, all colors are used. Clearly, there exists an edge x(i0)p0x(j0)q0∈E(˜G) such that β4(x(i0)p0x(j0)q0)=1, since in the coloring α2 there exists a vertex ui0 such that τi0(α2)=1, and since β4(x(i0)1x(j0)1)=(τi0(α2)−1)n+1. Similarly, there exists an edge x(i1)p1x(j1)q1∈E(˜G) such that β4(x(i1)p1x(j1)q1)=(W(G)+1)n−2, since in the coloring α2 there exists an edge ui1uj1∈E(G) with α2(ui1uj1)=W(G), and since β4(x(i1)n−1x(j1)n)=(α2(ui1uj1)−1)n+2n−2=(W(G)+1)n−2.
Now, by Lemma 1.1, we have that β4 is an interval ((W(G)+1)n−2)-edge coloring of the graph ˜G.
From Theorem 2.1, one can easily see that if (n)2=0 and Hi≅¯kn, then Theorem 2.1 can obtain the same upper bound on w(˜G) as Theorem 1.1.
We extend the result of Theorem 1.2 from the lexicographic product of graphs to the generalized lexicographic product of graphs and obtain the following result:
Theorem 2.2. If ˜G∈G2, then ˜G∈N. Moreover, w(˜G)≤w(G)n+w(Hk0), W(˜G)≥W(G)n+w(Hk0).
Proof. For the proof, let α be an interval t-coloring of G, and we consider the following two cases.
Case 1. ˜G∈G12.
Then Condition I holds, and by Lemma 1.2, we can assume that α(ui0uj0)=1.
Let
C1={w(Hk0)+1,…,w(Hk0)+tn};C2={1,…,w(Hk0)}; |
C(i)3={w(Hk0)−(w(Hi)−1),…,w(Hk0)},i∈{1,…,m}∖{k0}. |
Obviously, C1∩C2=∅, C(i)3⊆C2, and |C(i)3|=w(Hi).
Now, we construct an edge coloring β of ˜G as follows.
First, we define the following edge-coloring β1 of the graph G(1) with nt colors in C1. For every edge x(i)px(j)q∈E(G(1)), 1≤p,q≤n, let
β1(x(i)px(j)q)={w(Hk0)+(α(uiuj)−1)n+(p+q−1)n,p+q≠n+1;w(Hk0)+α(uiuj)n,p+q=n+1. |
Second, we define the following edge-coloring β2 of the graph G(2). If Hi≅Hk0, we color the edges of Hk0 with w(Hk0) colors in C2 such that the colors on the edges incident to any vertex are consecutive; if Hi≇Hk0, we color the edges of Hi with w(Hi) colors in C(i)3 such that the colors on the edges incident to any vertex are consecutive.
Finally, for every edge e∈E(˜G), let
β(e)={β1(e),e∈E(G(1));β2(e),e∈E(G(2)). |
Let us prove that β is an interval (tn+w(Hk0))-edge coloring of ˜G. It is easy to see that the set S(x(i)p,β) is an interval for each vertex x(i)p∈V(˜G), since both S(x(i)p,β1) and S(x(i)p,β2) are intervals, and since minS(x(i)p,β1)=maxS(x(i)p,β2)+1.
Next, we show that in the coloring β all colors are used. Clearly, there exists an edge x(i0)p0x(j0)q0∈E(˜G) such that β(x(i0)p0x(j0)q0)=1, since in the coloring α there exists an edge ui0uj0∈E(G) with α(ui0uj0)=1, and since β(x(i0)1x(j0)1)=(α(ui0uj0)−1)n+1. Similarly, there exists an edge x(i1)p1x(j1)q1∈E(˜G) such that β(x(i1)p1x(j1)q1)=tn+w(Hk0), since in the coloring α there exists an edge ui1uj1∈E(G) with α(ui1uj1)=t, and since β(x(i1)1x(j1)n)=α(ui1uj1)n+w(Hk0)=tn+w(Hk0).
Therefore, β is an interval (tn+w(Hk0)-edge coloring of ˜G. By the definition of β, we have w(˜G)≤w(G)n+w(Hk0), W(˜G)≥W(G)n+w(Hk0).
Case 2. ˜G∈G22.
Let
D1={1,…,tn};D2={tn+1,…,tn+w(Hk0)}; |
D(i)3={tn+w(Hk0)−(w(Hi)−1),…,tn+w(Hk0)},i∈{1,…,m}∖{k0}. |
Obviously, D1∩D2=∅, D(i)3⊆D2, and |D(i)3|=w(Hi).
Since G∈N, there exists an interval t-coloring α of G. Now, we construct an edge coloring β′ of ˜G as follows.
First, we define the following edge-coloring β3 of the graph G(1) with nt colors in D1. For every edge x(i)px(j)q∈E(G(1)), 1≤p,q≤n, let
β3(x(i)px(j)q)={(α(uiuj)−1)n+(p+q−1)n,p+q≠n+1;α(uiuj)n,p+q=n+1. |
Second, we define the following edge-coloring β4 of the graph G(2). If Hi≅Hk0, we color the edges of Hk0 with w(Hk0) colors in D2 such that the colors on the edges incident to any vertex are consecutive; if Hi≇Hk0, we color the edges of Hi with w(Hi) colors in D(i)3 such that on the edges incident to any vertex are consecutive.
Finally, for every edge e∈E(˜G), let
β′(e)={β3(e),e∈E(G(1));β4(e),e∈E(G(2)). |
It is easy to see that β′ is an interval (tn+w(Hk0))-edge coloring of the graph ˜G. Its proof is similar to the proof of case 1. By the definition of β′, we have w(˜G)≤w(G)n+w(Hk0), W(˜G)≥W(G)n+w(Hk0).
It is not difficult to see that if Hi is a r-regular graph and Hi∈N for any i=1,…,m, from Theorem 2.2, we can directly derive Theorem 1.2.
We extend Theorem 1.3 from the lexicographic product of graphs to the generalized lexicographic product of graphs, and obtained the following results:
Theorem 2.3. If ˜G∈G3, then ˜G∈N. Moreover,
(i) If ˜G∈G13, then w(˜G)≤w(G)n+2, W(˜G)≥W(G)n+n2+1;
(ii) If ˜G∈G23, then w(˜G)≤w(G)n, W(˜G)≥W(G)n+n2−1.
Proof. For the proof, we consider the following two cases:
Case 1. ˜G∈G13.
Then Condition I holds, and by Lemma 1.2, we can assume that α(ui0uj0)=1. Now, we prove that w(˜G)≤w(G)n+2. For this, we define an edge-coloring β1 of the graph ˜G in the following two steps:
Step 1. For every edge x(i)px(j)q∈E(G(1)), 1≤p,q≤n, let
β1(x(i)px(j)q)={(α1(uiuj)−1)n+(p+q−1)n+2,p+q≠n+1;α1(uiuj)n+2,p+q=n+1. |
Step 2. For every edge x(i)px(i)p+1∈E(G(2)), 1≤p≤n−1, let
β1(x(i)px(i)p+1)=(τi(α1)−1)n+(p)2+1; |
if Hi is isomorphic to a cycle, for the edge (x(i)1x(i)n)∈E(G(2)), let
β1(x(i)1x(i)n)=(τi(α1)−1)n+1. |
Let us prove that β1 is an interval edge coloring of the graph ˜G.
First, we prove that the set S(x(i)p,β1) is an interval for each vertex x(i)p∈V(˜G), where i=1,…,m, p=1,…,n.
If Hi≅¯Kn, by the definition of β1, we have
S(x(i)p,β1)={(τi(α1)−1)n+3,…,τ′i(α1)n+2}. |
If Hi is isomorphic to a path or a cycle, by the definition of β1, we have
S(x(i)p,β1)={(τi(α1)−1)n+3,…,τ′i(α1)n+2}∪{(τi(α1)−1)n+(p−1)2+1}∪{(τi(α1)−1)n+(p)2+1=[(τi(α1)−1)n+1,τ′i(α1)n+2],2≤p≤n−1; |
and when Hi is isomorphic to a path, we have
S(x(i)1,β1)=S(x(i)n,β1)={(τi(α1)−1)n+3,…,τ′i(α1)n+2}∪{(τi(α1)−1)n+1+1}=[(τi(α1)−1)n+2,τ′i(α1)n+2], |
when Hi is isomorphic to a cycle, we have
S(x(i)1,β1)=S(x(i)n,β1)={(τi(α1)−1)n+3,…,τ′i(α1)n+2}∪{(τi(α1)−1)n+1}∪{(τi(α1)−1)n+1+1}=[(τi(α1)−1)n+1,τ′i(α1)n+2]. |
Second, note that
d˜G(x(i)p)=dH(x(i)p)+dG(x(i)p)n={(τ′i(α1)−τi(α1))n+n,Hi≅¯Kn;(τ′i(α1)−τi(α1))n+n+1,Hi≅Pnandp=1orp=n;(τ′i(α1)−τi(α1))n+n+2,Hi≅Cn,orHi≅Pnand2≤p≤n−1. |
Clearly, we have
maxS(x(i)p,β1)−minS(x(i)p,β1)=d˜G(x(i)p)−1,1≤p≤n. |
This implies that β1 is a proper edge coloring of ˜G.
Finally, we show that β1 all colors are used. Clearly, there exists an edge x(i0)p0x(i0)p0+1∈E(˜G) such that β1(x(i0)p0x(i0)p0+1)=1, since in the coloring α1 there exists a vertex ui0 such that τi0(α1)=1, and since β1(x(i0)1x(i0)2)=(τi0(α1)−1)n+1=1. Similarly, there exists an edge x(i1)p1x(j1)q1∈E(˜G) such that β1(x(i1)p1x(j1)q1)=w(G)n+2, since in the coloring α1 there exists an edge ui1uj1∈E(G) with α(ui1uj1)=w(G), and since β1(x(i1)1x(j1)n)=α(ui1uj1)n+2=w(G)n+2.
Now, by Lemma 1, we have that β1 is an interval (w(G)n+2)-edge coloring of the graph ˜G.
Now, we prove that W(˜G)≥W(G)n+n2+1. For this, we define an edge-coloring β2 of the graph ˜G in the following two steps:
Step 1. For every edge x(i)px(j)q∈E(G(1)), let
β2(x(i)px(j)q)={(α2(uiuj)−1)n+p+q+1,1≤p≤n2and1≤q≤n2;(α2(uiuj)+1)n+3−p−q,n2+1≤p≤nandn2+1≤q≤n;α2(uiuj)n+n2+2−|p−q|,otherwise; |
Step 2. For every edge x(i)px(i)p+1∈E(G(2)), 1≤p≤n−1, let
β1(x(i)px(i)p+1)=β1(x(i)n−px(i)n−p+1)=(τi(α2)−1)n+p+1,1≤p≤n2; |
if Hi is isomorphic to a cycle, for the edge (x(i)1x(i)n)∈E(G(2)), let
β1(x(i)1x(i)n)=(τi(α2)−1)n+1. |
Let us prove that β2 is an interval (W(G)n+n2+1)-edge coloring of the graph ˜G.
First we prove that the set S(x(i)p,β2) is an interval for each vertex x(i)p∈V(˜G), where i=1,…,m, p=1,…,n.
If Hi is isomorphic to a path or a cycle, by the definition of β2, we have
S(x(i)p,β1)=S(x(i)n−p+1,β1)={(τi(α2)−1)n+p+2,…,τ′i(α2)n+n2+2−|p−n2−1|}∪{(τi(α2)−1)n+p}∪{(τi(α2)−1)n+p+1}=[(τi(α2)−1)n+p,τ′i(α2)n+p+1],2≤p≤n2; |
and when Hi is isomorphic to a path, we have
S(x(i)1,β1)=S(x(i)n,β1)={(τi(α2)−1)n+3,…,τ′i(α2)n+2}∪{(τi(α2)−1)n+2}=[(τi(α2)−1)n+2,τ′i(α2)n+2]; |
when Hi is isomorphic to a cycle, we have
S(x(i)1,β1)=S(x(i)n,β1)={(τi(α2)−1)n+3,…,τ′i(α2)n+2}∪{(τi(α2)−1)n+2}∪{(τi(α2)−1)n+1}=[(τi(α2)−1)n+1,τ′i(α2)n+2]; |
If Hi≅¯Kn, by the definition of β2, we have
S(x(i)p,β1)=S(x(i)n−p+1,β1)={(τi(α2)−1)n+p+2,…,τ′i(α2)n+p+1},1≤p≤n2; |
Second, note that
d˜G(x(i)p)=dH(x(i)p)+dG(x(i)p)n={(τ′i(α2)−τi(α2))n+n,Hi≅¯Kn;(τ′i(α2)−τi(α2))n+n+1,Hi≅Pnandp=1orp=n;(τ′i(α2)−τi(α2))n+n+2,Hi≅Cn,orHi≅Pnand2≤p≤n−1. |
Clearly, we have
maxS(x(i)p,β2)−minS(x(i)p,β2)=d˜G(x(i)p)−1,1≤p≤n. |
This implies that β2 is a proper edge coloring of ˜G.
Finally, we show that in the coloring β2, all colors are used. Clearly, there exists an edge x(i0)1x(i0)n∈E(˜G) such that β1(x(i0)1x(i0)n)=1, since in the coloring α2 there exists a vertex ui0 such that τi0(α2)=1, and since β2(x(i0)1x(i0)n)=(τi0(α2)−1)n+1. Similarly, there exists an edge x(i1)p1x(j1)q1∈E(˜G) such that β2(x(i1)p1x(j1)q1)=W(G)n+n2+1, since in the coloring α2 there exists an edge ui1uj1∈E(G) with α2(ui1uj1)=W(G), and since β2(x(i1)p1x(j1)q1)=(W(G)−1)n+3n2+1=W(G)n+n2+1.
Now, by Lemma 1.1, we have that β2 is an interval (W(G)n+n2+1)-edge coloring of the graph ˜G.
Case 2. ˜G∈G23.
Now, we prove that w(˜G)≤w(G)n. For this, we define an edge-coloring β3 of the graph ˜G in the following two steps:
Step 1. For every edge x(i)px(j)q∈E(G(1)), let
β3(x(i)px(j)q)={(α1(uiuj)−1)n+(p+q−1)n,p+q≠n+1;α1(uiuj)n,p+q=n+1. |
Step 2. For every edge x(i)px(i)p+1∈E(G(2)), 1≤p≤n−1, let
β3(x(i)px(i)p+1)=(τi(α1)−1)n−(p+1)2; |
if Hi is isomorphic to a cycle, for the edge (x(i)1x(i)n)∈E(G(2)), let
β3(x(i)1x(i)n)=(τi(α1)−1)n−1. |
Similar to the coloring β1 in case 1 to discuss. It is easy to see that β3 is an interval w(G)n-coloring of ˜G.
Now, we prove that W(˜G)≥W(G)n+n2−1. For this, we define an edge-coloring β4 of the graph ˜G in the following two steps:
Step 1. For every edge x(i)px(j)q∈E(G(1)), 1≤p,q≤n, let
β4(x(i)px(j)q)=β2(x(i)px(j)q)−2; |
Step 2. For every edge x(i)px(i)q∈E(G(2)), 1≤p,q≤n, let
β4(x(i)px(i)q)=β2(x(i)px(i)q). |
It is easy to see that β4 is an interval (W(G)n+n2−1)-coloring of ˜G.
It is not difficult to see that if Hi is isomorphic to a cycle for any i=1,…,m, from Theorem 2.3, we can directly derive Theorem 1.3.
From Theorems 2.1–2.3, we can see that Gz⊂N,z=1,2,3.
In this paper, we studied the interval edge coloring of the generalized lexicographic product ˜G of an interval colorable graph G with m vertices and a sequence of vertex-disjoint graphs hm=(Hi)i∈{1,…,m}, where each graph in hm has n vertices, and proved that ˜G is interval colorable if and only if hm=(Hi)i∈{1,…,m} satisfies one of the following three conditions: (i) each graph Hi in hm is isomorphic to a path or an empty graph; (ii) each graph Hi in hm is isomorphic to an empty graph or an interval-colorable regular graph, but not all graphs in hm are empty graphs; (iii) each graph Hi in hm is isomorphic to a path or a cycle or an empty graph of even order, and Hk0 is a cycle. Moreover, we obtain the bounds on w(˜G) and W(˜G).
M. Jin: Writing-original draft preparation; M. Jin, P. Chen and S. Tian: Formal analysis; M. Jin, P. Chen and S. Tian: Writing-review and editing. All authors have read and agreed to the published version of the manuscript.
This research was supported by the National Natural Science Foundation of China (12061061) and Applied Mathematics National Minority Committee Key Discipline (11080327). The authors thank the innovation team of operations research and cybernetics at Northwest Minzu University.
All authors declare no conflicts of interest in this paper.
[1] |
A. S. Asratian, R. R. Kamalian, Interval colorings of edges of a multigraph, Appl. Math., 5 (1987), 25–34. https://doi.org/10.48550/arXiv.1401.8079 doi: 10.48550/arXiv.1401.8079
![]() |
[2] |
A. S. Asratian, C. J. Casselgren, On interval edge colorings of (α,β)-biregular bipartite graphs, Discrete Math., 307 (2007), 1951–1956. https://doi.org/10.1016/j.disc.2006.11.001 doi: 10.1016/j.disc.2006.11.001
![]() |
[3] | J. A. Bondy, U. S. R. Murty, Graph theory with applications, London: Macmillan, 1976. |
[4] |
K. Giaro, M. Kubale, M. Małafiejski, Consecutive colorings of the edges of general graphs, Discrete Math., 236 (2001), 131–143. https://doi.org/10.1016/S0012-365X(00)00437-4 doi: 10.1016/S0012-365X(00)00437-4
![]() |
[5] |
A. Grzesik, H. Khachatrian, Interval edge-colorings of K1,m,n, Discrete Appl. Math., 174 (2014), 140–145. https://doi.org/10.1016/j.dam.2014.04.003 doi: 10.1016/j.dam.2014.04.003
![]() |
[6] | H. M. Hansen, Scheduling with minimum waiting periods, Master's Thesis, Odense University, Denmark, 1992. |
[7] |
I. Holyer, The NP-completeness of edge-coloring, SIAM J. Comput., 10 (1981), 718–720. https://doi.org/10.1137/0210055 doi: 10.1137/0210055
![]() |
[8] |
P. Jing, Z. Miao, Z. X. Song, Some remarks on interval colorings of complete tripartite and biregular graphs, Discrete Appl. Math., 277 (2020), 193–197. https://doi.org/10.1016/j.dam.2019.08.024 doi: 10.1016/j.dam.2019.08.024
![]() |
[9] | R. R. Kamalian, Interval colorings of complete bipartite graphs and trees, Preprint of the Computing Centre of the Academy of Sciences of Armenia, 1989. https://doi.org/10.48550/arXiv.1308.2541 |
[10] | R. R. Kamalian, Interval edge colorings of graphs, Doctoral Thesis, Novosibirsk, 1990. |
[11] | R. R. Kamalian, P. A. Petrosyan, A note on upper bounds for the maximum span in interval edge-colorings of graphs, Discrete Math., 312 (2012), 1393–1399. https://doi.org/10.1016/j.disc.2012.01.005 |
[12] | P. A. Petrosyan, Interval edge-colorings of Möbius ladders, In: Proceedings of the CSIT Conference, 2005,146–149. https://doi.org/10.48550/arXiv.0801.0159 |
[13] | P. A. Petrosyan, Interval edge-colorings of complete graphs and n-dimensional cubes, Discrete Math., 310 (2010), 1580–1587. doi:10.1016/j.disc.2010.02.001 |
[14] | P. A. Petrosyan, Interval edge colorings of some products of graphs, Discuss. Math. Graph T., 31 (2011), 357–373. |
[15] | P. A. Petrosyan, H. H. Khachatrian, L. E. Yepremyan, H. G. Tananyan, Interval edge-colorings of graph products, In: Proceedings of the CSIT Conference, 2011, 89–92. https://doi.org/10.48550/arXiv.1110.1165 |
[16] | S. V. Sevastjanov, Interval colorability of the edges of a bipartite graph, Metody Diskretnogo Analiza, 50 (1990), 61–72. |
[17] |
V. Samodivkin, Domination related parameters in the generalized lexicographic product of graphs, Discrete Appl. Math., 300 (2021), 77-84. https://doi.org/10.1016/j.dam.2021.03.015 doi: 10.1016/j.dam.2021.03.015
![]() |
[18] |
H. H. Tepanyan, P. A. Petrosyan, Interval edge-colorings of composition of graphs, Discrete Appl. Math., 217 (2017), 368–374. https://doi.org/10.1016/j.dam.2016.09.022 doi: 10.1016/j.dam.2016.09.022
![]() |
[19] | V. G. Vizing, On an estimate of the chromatic class of a p-graph, Discret Analiz, 3 (1964), 25–30. |