e-injective coloring: 2-distance and injective coloring conjectures

Shahrzad. S. Mirdamad1, Doost Ali Mojdeh1
1Department of Mathematics, Faculty of Mathematical Sciences, University of Mazandaran, Babolsar, Iran

Abstract

An injective coloring of a given graph G=(V,E) is a vertex coloring of G such that any two vertices with a common neighbor receive distinct colors. An e-injective coloring of a graph G is a vertex coloring of G in which any two vertices v,u with a common edge e (euv) receive distinct colors; in other words, any two end vertices of a path P4 in G achieve different colors. With this new definition, we want to take a review of injective coloring of a graph from the new point of view. For this purpose, we review the conjectures raised so far in the literature of injective coloring and 2-distance coloring, from the new approach of e-injective coloring. Additionally, we prove that, for disjoint graphs G,H, with E(G) and E(H), χei(GH)=max{χei(G),χei(H)} and χei(GH)=|V(G)|+|V(H)|. The e-injective chromatic number of G versus the maximum degree and packing number of G is investigated, and we denote max{χei(G),χei(H)}χei(G◻H)χ2(G)χ2(H). Finally, we prove that, for any tree T (T is not a star), χei(T)=χ(T), and we obtain the exact value of the e-injective chromatic number for some specified graphs.

Keywords: Injective coloring conjecture, 2-distance coloring conjecture, e-injective coloring

1. Introduction

Graph coloring has many applications in various fields of life, such as timetabling, scheduling daily life activities, scheduling computer processes, registering allocations to different institutions and libraries, manufacturing tools, printed circuit testing, routing and wavelength, bag rationalization for a food manufacturer, satellite range scheduling, and frequency assignment. These are many applications that are out there right now and many more come in the follow.

A proper k-coloring (hereafter k-coloring) of a graph G is a function f:V(G){1,2,3,,k} such that for all edge xyE, f(x)f(y). The chromatic number of G, denoted by χ(G), is the minimum integer k such that G has a k-coloring. There are many research works on the graph coloring parameters that is not possible to name all of them here. For instance see [4,17].

In 1969, Kramer and Kramer introduced the notion of 2distance coloring of a graph G or the usual proper coloring of G2 [9], we can some of its applications in [10]. A 2distance k-coloring of a graph G is a function f:V{1,2,3,,k}, such that no pair of vertices at distance at most 2, receive the same color, in the other words, the colors of the vertices of all P3 paths in the graph are distinct. The 2-distance chromatic number of G, denoted by χ2(G)=χ(G2), is the minimum positive integer k such that G has a 2-distance k-coloring. The 2-distance coloring of G, is a proper coloring [3,2,5].

For a graph G, the subset S of V(G) is said to be a dominating set if any vertex xVS is adjacent to a vertex y in S. A dominating set of G with minimum cardinality is called the domination number of G and is denoted by γ(G). A subset D of V(G) is said to be 2-distance dominating set if any vertex dVD, is in at most 2-distance of to a vertex in D. A 2-distance dominating set of G with minimum cardinality is called the 2-distance domination number of G and is denoted by γ2(G).

The injective coloring was first introduced in 2002 by Hahn et al. [6] and it was also further studied in [1,8,11,12,15,16,18]. An injective k-coloring of a graph G is a function f:V{1,2,3,,k} such that no vertex v is adjacent to two vertices u and w with f(u)=f(w), in the other words, for any path P3=xyz, we have f(x)f(z). The injective chromatic number of G, denoted by χi(G), is the minimum positive integer k such that G has an injective k-coloring. The injective chromatic number of the hypercube has important applications in the theory of error-correcting codes. As it is well known, the injective coloring of G, is not necessarily proper coloring. Injective coloring of a graph G is related to the usual coloring of the square G2. The inequality χi(G)χ2(G) obviously holds.

There are several results related to injective coloring that review the usual coloring results in graph theory from a new point of view, in particular on the injective chromatic number of planar graphs. As well, many conjectures on planar graphs have been posed and studied by authors so far. In this regard, we bring up some of them as follows.

From the relation between the injective coloring of a graph G and the usual coloring of the square G2, Wegner [19] posed the following conjecture.

Conjecture 1.1. [19] Let G be a planar graph with maximum degree Δ. Then,

1. χ(G2)7  if Δ=3,

2. χ(G2)Δ+5  if 4Δ7,

3. χ(G2)32Δ+1  otherwise.

Lužar and Škrekovski in [11] showed that:

Theorem 1.2. ([11] Theorem 2.1) There exist planar graphs G of maximum degree Δ3 satisfying the following,

1. χi(G)=5  if Δ=3,

2. χi(G)=Δ+5  if 4Δ7,

3. χi(G)=32Δ+1  if Δ8.

Adapted from Theorem 1.2, they proposed the following Wegner type conjecture for the injective chromatic number of planar graphs.

Conjecture 1.3. [11] Let G be a planar graph with maximum degree Δ. Then,

(i). χi(G)5  if Δ=3,

(ii). χi(G)Δ+5  if 4Δ7,

(iii). χi(G)32Δ+1  otherwise.

By the relation between injective chromatic number and 2-distance chromatic number of a graph; showing the truth of Conjecture 1.1 (parts (2), (3)), will deduce the truth of Conjecture 1.3 (parts (ii), (iii)).

Now we introduce a new concept of vertex coloring (near to injective coloring) as an e-injective coloring of a graph. The motivation of the alleging is to study, how it behaves against of the injective graph coloring, usual graph coloring, 2-distance graph coloring, packing set, dominating set and 2-distance dominating set of graphs. As well, in particular we investigate the posed conjectures from the point of view of e-injective colorings. Also since the notion of e-injective coloring is near to injective coloring, one can predict, it has applications in various fields of life in real world and would also be useful in coding theory as so did injective coloring.

This concept is introduced in next definition.

Hereafter, we say that, two vertices u,v have a common edge neighbor if there exist an edge e in which, u is adjacent to one end vertex of e and v is adjacent to another end vertex of e.

Definition 1.4. Let G be a graph. A function f:V(G){1,2,3,,k} is an e-injective k-coloring function if any two vertices u and v are the ends of a path P4=uxyv in G, then f(u)f(v).

The einjective chromatic number of G, denoted by χei(G), is the minimum positive integer k such that G has an e-injective k-coloring.

The e-injective coloring of G, is not necessarily proper coloring. This concept can be expressed as new discourse.

Definition 1.5. For a given graph G, the three-step graph S3(G)=G(3) of a graph G is the graph having the same vertex set as G with an edge joining two vertices in S3(G) if and only if there is a path of length 3 between them in G.

Taking into account, the fact that a vertex subset S is independent in S3(G) if and only if there is no path of length 3 between any two vertices corresponding of S in G, we can readily observe that: χei(G)=χ(S3(G)).

From the point of view of e-injective coloring, the type of Conjectures 1.1 and 1.3 can be declared as a problem, which can be argued in next section.

Problem 1.6. Let G be a planar graph with maximum degree Δ. Then,

1. χei(G)5  if Δ=3,

2. χei(G)Δ+5  if 4Δ7,

3. χei(G)32Δ+1  otherwise.

In the sequence, we assume that all graphs in this paper are finite, simple, and undirected. We use [4,20] as a reference for terminology and notation which are not explicitly defined here. Throughout the paper, we consider G=(V,E) be a finite simple graph with vertex set V=V(G) and edge set E=E(G). The open neighborhood of a vertex vV is the set N(v)={u|uvE}, and its closed neighborhood is the set N[v]=N(v){v}. The cardinality of |N(v)| is called the degree of v, denoted by deg(v). The minimum degree of G is denoted by δ(G) and the maximum degree by Δ(G). A vertex v of degree 1 is called a pendant vertex or a leaf, and its neighbor is called a support vertex. A vertex of degree n1 is called a full or universal vertex while a vertex of degree 0 is called an isolated vertex.

For any two vertices u and v of G, we denote by dG(u,v) the distance between u and v, that is the length of a shortest path joining u and v. The path, cycle and complete graph with n vertices are denoted by Pn, Cn and Kn respectively. The complete bipartite graph with n and m vertices in their partite sets is denoted by Kn,m, while the wheel graph with n+1 vertices is denoted by Wn. A star graph with n+1 vertices, denoted by Sn, consists of n leaves and one support vertex. A double star graph is a graph consisting of the union of two star graphs Sm and Sn, with one edge joining their support vertices; the double star graph with m+n+2 vertices is denoted by Sm,n.

The join of two graphs G and H, denoted by GH, is the graph obtained from the disjoint union of G and H with vertex set V(GH)=V(G)V(H) and edge set E(GH)=E(G)E(H){xy | xV(G), yV(H)}. A fan graph is a simple graph consisting of joining K¯m and Pn; the fan graph with m+n vertices is denoted by Fm,n. For two sets of vertices X and Y, the set [X,Y] denotes the set of edges e=uv such that uX and vY.

The square graph G2 is a graph with the same vertex set as G and with its edge set given by E(G2)={uv| dist(u,v)2}. The chromatic number χ(G2) of G2 (or 2-distance chromatic number χ2(G) of G) has been studied extensively in planar graph [7,14].

A subset BV(G) is a packing set in G if for every pair of distinct vertices u,vB, NG[u]NG[v]=. The packing number ρ(G) is the maximum cardinality of a packing set in G.
A subset BV(G) is an open packing set in G if for every pair of distinct vertices u,vB, NG(u)NG(v)=. The open packing number ρ(G) is the maximum cardinality among all open packing sets in G.

Let G and H be simple graphs. For three standard products of graphs G and H, the vertex set of the product is V(G)×V(H) and their edge set is defined as follows:

  • In the Cartesian product G◻H, two vertices are adjacent if they are adjacent in one coordinate and equal in the other.

  • In the direct product G×H, two vertices are adjacent if they are adjacent in both coordinates.

  • The edge set of the strong product GH, is the union of E(G◻H) and E(G×H).

In the end of this section, we explore the purpose of the paper as follows. In Section 2, we study χei(G) versus to the χ(G), χ2(G) and χi(G), as well, we review the conjectures raised so far in the literature of injective and 2-distance colorings, from the new approach, e-injective coloring, and by disproving the Problem 1.6, we show that the Conjectures 1.1 and 1.3 maybe wrong under the conditions. For disjoint graphs G,H, with non-empty edge sets, χei(GH)=max{χei(G),χei(H)} and χei(GH)=|V(G)|+|V(H)|. The e-injective chromatic number of G versus of the maximum degree and packing number of G is investigated, and denote max{χei(G),χei(H)}χei(G◻H)χ2(G)χ2(H) in Section 3. In Section 4, we prove that, for any tree T (TSn), χei(T)=χ(T), and we obtain the exact value of e-injective chromatic number of some special graphs and finally, we end the paper with discussion on research problems.

2. On the two conjectures

We maybe cannot compare χei(G) with χ(G), Δ(G), χi(G) or χ2(G) in general. For instance, χei(K3)=1 while χ(K3)=3=χi(K3)=χ2(G); or χei(K1,n)=1 while χ(K1,n)=2, χi(K1,n)=n and χ2(K1,n)=n+1. But in the Figure 1, χei(G)=12, χi(G)7, χ2(G)7, χ(G)=3.

Figure 1 The graph G with χei(G)max{χi(G),χ2(G),χ(G)}

Also let H=KmKn be a graph obtain from two complete graphs Km and Kn (mn4) with joining one vertex of Km to one vertex of Kn. Then χ(H)=m=χi(H), χ2(H)=m+1 and χei(H)=m+n1 (see Proposition 4.4). On the other hand, for Complete graph Kn (n4), odd cycle C3k for odd k1, we have χei(Kn)=n=χ(Kn)=χi(Kn)=χ2(Kn), and χei(C3k)=3=χ(C3k)=χi(C3k)=χ2(C3k).

In the same way, we have Δ(K3)=2, Δ(H)=m, and for graph G in Figure 1, Δ(G)=4, while χei(K3)=1, χei(H)=m+n1, and χei(G)=12. On the other hand, for graph KnK1 (n4) and even cycle C2k, we have χei(KnK1)=n=Δ(KnK1), and χei(C2k)=2=Δ(C2k) (see Proposition 4.3).

However we have the following.

Proposition 2.1. Let G be a graph in which any two adjacent vertices be the end vertices of a path P4 in G. Then χ(G)χei(G).

Conversely, if any two end vertices of each path P4 in G are adjacent, then χei(G)χ(G).

Proof. Since any two adjacent vertices of given graph G are the end vertices of a path P4 in G, these two vertices must be colored with distinct colors by any e-injective coloring. Therefore, any e-injective coloring for this graph can be a usual coloring. Then χ(G)χei(G).

Conversely, by the construction of graph G, usual coloring of G deduces that any two end vertices of each path P4 has different colors. Thus, the usual coloring of given G is an e-injective coloring. Therefore, χei(G)χ(G)◻

Proposition 2.2. Let G be a graph and v in V(G) be any vertex. If every two vertices in N(v) are the end vertices of a path P4, then χi(G)χei(G).

Conversely, if end vertices of each path P4 in a graph G have a common neighbor, then χei(G)χi(G).

Proof. Since any two vertices of graph G are the end vertices of a path P4, these vertices have distinct colors by e-injective coloring. Therefore, an e-injective coloring for this graph is an injective coloring. Then χi(G)χei(G).

Conversely, let any two end vertices of path P4 have a common neighbor. Then by injective coloring of G, two end vertices of any path P4 take different colors. Thus, this will be an e-injective coloring of G and χei(G)χi(G)◻

Also, we may have.

Proposition 2.3. Let G be a graph with the property that, for any two adjacent vertices or two vertices with a common neighbor are the end vertices of a path P4 in G. Then χ2(G)χei(G).

Conversely, if G is a graph and any two end vertices of each path P4 in G are adjacent or have a common neighbor, then χei(G)χ2(G).

Proof. Let v,u,w be three vertices of P3 in G. Since both of them are the end vertices of a path P4 in G, then e-injective coloring of the given graph G assign three distinct colors to v,u,w. This implies that, this coloring is a 2-distance coloring of G. Thus, χ2(G)χei(G).

Conversely, let any two end vertices of each path P4 are adjacent or have a common neighbor. Then, from a 2-distance coloring of G, we deduce that, any two end vertices of the path P4 are vertices of a P2 or a P3 in graph G and so their colors are distinct. Therefore this coloring is an e-injective coloring of the given graph G and χei(G)χ2(G)◻

Now we discuss on the Problem 1.6. Below figures denote that the Problem 1.6 is not necessarily true. On the other hand the type of Conjectures 1.1, 1.3 are not true for e-injective coloring. But if we use the Propositions 2.2, 2.3, then maybe characterize graphs G in which, satisfy on the Conjectures 1.1, 1.3 and also characterize graphs G in which, the Conjectures 1.1, 1.3 are not true for them.

Disprove of Problem 1.6
Let G be a planar graph with maximum degree Δ. Then we present counterexample that denote the Problem 1.6 is not true.

Figure 2 The graphs G related to Problem 1.6 for Δ3

As we observe the Figure 2, for (3Δ8) we have.

The graph M3 denotes a planar graph in which Δ(M3)=3 and χei(M3)=6>5.

The graph M4 denotes a planar graph in which Δ(M4)=4 and χei(M4)=12>Δ(M4)+5.

The graph M5 denotes a planar graph in which Δ(M5)=5 and χei(M5)=13>Δ(M5)+5.

The graph M6 denotes a planar graph in which Δ(M6)=6 and χei(M6)=16>Δ(M6)+5.

The graph M7 denotes a planar graph in which Δ(M7)=7 and χei(M7)=16>Δ(M7)+5.

For Δ(G)=8, consider the graph M8 of order 16, which is seen Δ(G)=8 and χei(Mn)=16>242+1=13=3Δ2+1.

For Δ(G)9 consider, the graph Mn (n9) of order 2n, Figure 2, which is seen Δ(G)=n and χei(Mn)=2n>3n2+1=3Δ2+1.

With this regard, the Problem 1.6 is disproved. In the other words the type of Conjecture 1.3 for e-injective coloring is rejected. However, from the Propositions 2.2 and 2.3, we can have.

Proposition 2.4. 1. Let G be a graph with the property that the given data in Proposition, 2.3 (Conversely part) hold. Then, the Conjecture 1.1 is wrong.

2. Let G be a graph with the property that the given data in Proposition, 2.2 (Conversely part) hold. Then, the Conjecture 1.3 is wrong.

3. e-injective chromatic number on operation of graphs

In this section we prove some results on e-injective coloring using some operations.

For graphs G and H, let GH be the disjoint union of G and H. Then it is easy to see that χei(GH)=max{χei(G),χei(H)}.

For the join of two graphs, we have the following.

Theorem 3.1. Let G and H be two graphs of order m and n respectively, with the property that, E(G) and E(H) are non-empty sets. Then χei(GH)=m+n

Proof. Let e1=v1w1E(G) and e2=v2w2E(H) be two edges. We show that any two vertices x,y in GH, there is a path of length 3, with end vertices x,y. For observing the result, we bring up five positions.

1. For x,yV(G), consider the path xv2w2y in GH.

2. For x,yV(H), consider the path xv1w1y in GH.

3. For xV(G){v1,w1} and yV(H){v2,w2}, consider the path xv2v1y in GH.

4. For x{v1,w1}, say v1 and yV(H){v2,w2}, consider the path v1v2w1y in GH.

5. For x{v1,w1} and y{v2,w2} and without loss of generality, say x=v1 and y=v2, consider the path v1w1w2v2 in GH.

The other positions are similar. Therefore, for any two vertices x,yGH there is a path of length 3, with end vertices x,y. Therefore the result is observed. ◻

Let G be a graph and B be a maximum packing set of G. If vV(G)B, then there is a vertex uB such that N(v)N(u). This shows that, d(v,u)2. Thus B is a 2-distance dominating set. Therefore we have.

Theorem 3.2. Let G be a graph of diameter 3. Then χei(G)ρ(G)γ2(G). One can have the equalities.

Proof. Let B be maximum packing set of graph G. Since two vertices of B has distance 3, they are assigned with two distinct colors. Thus χei(G)ρ.
For equalities, consider the cycles C6 and C8, (see Propositions 4.3). ◻

We now give an upper bound on χei(G) that may be slightly important.

Theorem 3.3. Let G have maximum degree Δ. Then, χei(G)Δ(Δ1)2+1. This bound is sharp for odd cycle Cn (n5).

Proof. Let G be a graph and vV(S3(G))=V(G). It is well known that there are at most Δ(Δ1)2 vertices in G such that any of them with v form two end vertices of path P4. This shows that degS3(G)(v)Δ(Δ1)2. On the other hand χei(G)=χ(S3(G)) and from Brooks Theorem in usual coloring of graphs, χ(S3(G))Δ(S3(G))+1Δ(Δ1)2+1. Therefore χei(G)Δ(Δ1)2+1.
For seeing the sharpness observe Proposition 4.3. ◻

Also we want to drive bounds for the e-injective coloring of Cartesian product of two graphs G, H in terms of 2-distance coloring of the of G and H. For this we explore a result from [13] and a lemma.

Theorem 3.4. ([13] Theorem 1) For any graphs G and H with no isolated vertices, (Δ(G)+1)(Δ(H)+1)χ2(GH)χ2(G)χ2(H).

Lemma 3.5. Let G and H be two graphs with no isolated vertices. If two end vertices of each path P4 in G and H are adjacent or have a common neighbor, then so does GH.

Proof. Suppose that the end vertices of each path P4 in graphs G and H are adjacent or have a common neighbor. We would to be show any two end vertices of a path P4 in GH are adjacent or have a common neighbor. For this, we can bring up the possible paths P4 in graph GH.
1.1. (a,u)(a,v)(a,w)(a,t);  1.2. (a,u)(a,v)(a,w)(b,w);  1.3. (a,u)(a,v)(a,w)(b,t).

2.1. (a,u)(a,v)(b,v)(b,w);  2.2. (a,u)(a,v)(b,v)(c,v);  2.3 (a,u)(a,v)(b,v)(c,w).

3.1. (a,u)(a,v)(b,w)(b,t);  3.2. (a,u)(a,v)(b,w)(c,w);  3.3. (a,u)(a,v)(b,w)(c,t).

4.1. (a,u)(b,u)(b,v)(b,w);  4.2. (a,u)(b,u)(b,v)(c,v);  4.3. (a,u)(b,u)(b,v)(c,w).

5.1. (a,u)(b,u)(c,u)(c,v);  5.2. (a,u)(b,u)(c,u)(d,u);  5.3. (a,u)(b,u)(c,u)(d,v).

6.1. (a,u)(b,u)(c,v)(c,w);  6.2. (a,u)(b,u)(c,v)(d,v);  6.3. (a,u)(b,u)(c,v)(d,w).

7.1. (a,u)(b,v)(b,w)(b,t);  7.2. (a,u)(b,v)(b,w)(c,w);  7.3. (a,u)(b,v)(b,w)(c,t).

8.1. (a,u)(b,v)(c,v)(c,w);  8.2. (a,u)(b,v)(c,v)(d,v);  8.3. (a,u)(b,v)(c,v)(d,w).

9.1. (a,u)(b,v)(c,w)(d,w);  9.2. (a,u)(b,v)(c,w)(c,t);  9.3. (a,u)(b,v)(c,w)(d,t).
Now we observe that, all these paths type P4 are adjacent or have a common neighbor.

1.1. Since uvwt is a path P4 in H, the vertices u and t are adjacent or have a common neighbor. If u and t are adjacent, then the vertices (a,u) and (a,t) are adjacent in GH. If u and t have a common neighbor s, then (a,s) is a common neighbor of (a,u) and (a,t).

1.2. (a,v) is a common neighbor of (a,u) and (b,w) in GH.

1.3. The uvwt is a path P4 in H. If u and t are adjacent, then the vertices (a,u) and (b,t) are adjacent in GH. If u and t have a common neighbor s, then (a,s) is a common neighbor of (a,u) and (b,t).

2.1. The vertex (b,v) is a common neighbor of (a,u) and (b,w) in GH.

2.2. The vertex (b,v) is a common neighbor of (a,u) and (c,v) in GH.

2.3. The vertex (b,v) is a common neighbor of (a,u) and (c,w) in GH.

3.1. Its proof is readily and similar to the proof of 1.3.

3.2. The vertex (b,v) is a common neighbor of (a,u) and (c,w) in GH.

3.3. Its proof is readily, and is similar to the proof of 1.3.

4.1. The vertex (b,v) is a common neighbor of (a,u) and (b,w) in GH.

4.2. The vertex (b,v) is a common neighbor of (a,u) and (c,v) in GH.

4.3. The vertex (b,v) is a common neighbor of (a,u) and (c,w) in GH.

5.1. The vertex (b,u) is a common neighbor of (a,u) and (c,v) in GH.

5.2. Since abcd is a path P4 in G, the vertices a and d are adjacent or have a common neighbor. If a and d are adjacent, then the vertices (a,u) and (d,u) are adjacent in GH. If a and d have a common neighbor r, then (r,u) is a common neighbor of (a,u) and (d,u).

5.3. Its proof is obvious and it is similar to the proof of 5.2.

6.1. The vertex (b,v) is a common neighbor of (a,u) and (c,v) in GH.

6.2. Its proof is obvious and it is similar to the proof of 5.2.

6.3. Its proof is obvious and it is similar to the proof of 5.2.

7.1. Its proof is similar to the proof of 1.3.

7.2. The vertex (b,v) is a common neighbor of (a,u) and (c,w) in GH.

7.3. Its proof is similar to the proof of 1.3.

8.1. The vertex (b,v) is a common neighbor of (a,u) and (c,w) in GH.

8.2. Its proof is similar to the proof of 5.2.

8.3. Its proof is similar to the proof of 5.2.

9.1. Its proof is similar to the proof of 5.2.

9.2. Its proof is similar to the proof of 1.3.

9.3. There are two paths abcd and uvwt in G and H respectively. If adE(G) and utE(H), then (a,d) and (u,t) are adjacent in GH. If adE(G) and s is a common neighbor of u and t, then (a,s) is a common neighbor of (a,u) and (d,t) in GH. If utE(H) and r is a common neighbor of a and d, then (r,u) is a common neighbor of (a,u) and (d,t) in GH. If a and d have a common neighbor r, and similarly, s is a common neighbor of u and t, then (r,s) is a common neighbor of (a,u) and (d,t) in GH.

It is observed that, both end vertices of every path P4 in GH are adjacent or have a common neighbor. Therefore the proof is complete. ◻

Now we have the following.

Theorem 3.6. For any graphs G and H with no isolated vertices, with the property that, any two end vertices of each path P4 in G and H are adjacent or have a common neighbor, we have Max{χei(G),χei(H)}χei(G◻H)χ2(G)χ2(H). The bounds are sharp.

Proof. For the first inequality, since G and H have no isolated vertices, and any path of length 3 of G and H gives at least one path of length 3 of G◻H, thus the first inequality holds. For seeing the sharpness, consider G=Pm and H=Pn where m4 or n4.

We now prove the second inequality. From the definitions of Cartesian and strong products, we may have G◻H as a subgraph of GH, and next any path P4 of (G◻H) is a path P4 of (GH). Therefore, χei(G◻H)χei(GH). As the same way, χ2(G◻H)χ2(GH). From Proposition 2.3 and Lemma 3.5 χei(GH)χ2(GH). On the other hand, from Theorem 3.4, χ2(GH)χ2(G)χ2(H). These deduce that χei(G◻H)χ2(G)χ2(H). It is easy to see that, this bound is sharp for G=C3 and H=C5 and also G=C3 and H=C7◻

4. e-injective chromatic number of special graphs

In this section we investigate the e-injective coloring of some special graphs, such as trees, path, cycle, complete graphs, wheel graphs, star, complete bipartite graphs, k-regular bipartite graphs, multipartite graphs and fan graphs.

Theorem 4.1. Let T be a tree. Then,

1. χei(T)=1 if and only if diam(T)2.

2. χei(T)=2 if and only if diam(T)3.

Proof. 1. If diam(T)=1, then T=P2 and if diam(T)=2, then T is a star and since there is no path of length 3 between any two vertices, χei(T)=1.

Conversely, let χei(T)=1. Then there is no path of length 3 in T. Thus diam(T)2.

2. Let diam(T)3 and v0 be a vertex of maximum degree in T. We assign color 1 to the v0 and to the vertex u if d(u,v0) is even, and color 2 to the vertex u if d(u,v0) is odd. Since there is only one path between any two vertices in any tree T, so if two vertices x,y are in distance 3 and two vertices x,z are in distance 3, then two vertices y,z are not in distance 3. This shows that, we can use color 1 for x and color 2 for y,z. Therefore χei(T)2. On the other hand, if diam(T)3, then χei(T)2. Therefore, if diam(T)3, then χei(T)=2.

Conversely, if χei(T)=2, there is two vertices v,w in T so that the path vxyw is of length 3 in T. This shows that diam(T)3◻

As an immediate from Theorem 4.1 we have.

Proposition 4.2. For Path Pn, we have

χei(Pn)={1  n3,2  n4.

Proposition 4.3. For cycle Cn, we have χei(Cn)={1n=3,2n4, 2|n,3n4, 2n.

Proof. Let n=3. It is obvious that χei(C3)=1.

Let n4. There are two cases to be considered.

Case 1. If n is even.

We assign the color 1 to the odd vertices and color 2 to the even vertices. Therefore χei(C2k)=2.

Case 2. If n is odd.

Let n=5. We assign the color 1 to the vertices v1,v2 and color 2 to the vertices v3,v4 and we assign color 3 to the vertex v5. This assignments is an e-injective coloring of C5.

Let n7 . We assign the color 1 to the odd vertices vis, for in4 and color 2 to the even vertices vis, for in3 and we assign color 3 to the vertices vn2,vn1,vn. This assignments is an e-injective coloring of Cn for odd n7. On the other hand, since there are two paths of length 3 between vn2 with two vertices v1, vn5, as well as vn1 with two vertices v2, vn4 and also vn with two vertices v3, vn3. It is clear that, one cannot e-injective color to the vertices cycle Cn withe two colors for odd n. Therefore the result holds. ◻

Since for n4, any two vertices of Kn are end of a path P4, then we have.

Proposition 4.4. For complete graph Kn, we have χei(Kn)={1  n3,n  n4.

Proposition 4.5. For wheel graph Wn(n3), χei(Wn)=n+1.

Proof. Let v1 be a universal vertex. For i,j2, there exists a path vjvj+1v1vi between vi and vj if vivj+1; and there exists a path vjv1vj+2vi between vi and vj if vi=vj+1. On the other hand, there exists a path v1vi2vi1vi between v1 and vi. Taking this account, there exist a path of length 3 between two vertices in Wn. Therefore χei(Wn)=n+1◻

Proposition 4.6. For complete bipartite graph Kn,m with m,n2, χei(Kn,m)=2.

Proof. Let n2, m2. It is easy to observe that, there is a path of length 3 between two vertices of two different partite sets. Therefore one can assign color 1 to one partite set and 2 to another partite set. Thus, χei(Kn,m)=2◻

Using the proof of Proposition 4.6, For regular bipartite graphs, we have the next result, which proof is similar.

Proposition 4.7. For k-regular bipartite graph G (k2), χei(G)=2.

A complete r partite graph is a simple graph such that the vertices are partitioned to r independent vertex sets and every pair of vertices are adjacent if and only if they belong to different partite sets.

Proposition 4.8. Let G=Kn1,,nr be a complete (r3) partite graph of order n. Then

χei(G)={1,  r=3  and  G=K1,1,1,n1,  r=3  and  G{Kn2,1,1,K1,n2,1,K1,1,n2}  with  n4,n,  otherwise.

Proof. 1. It is trivial.

2. Let r=3 and G{Kn2,1,1,K1,n2,1,K1,1,n2}. Without loss of generality, assume that G=Kn2,1,1 with vertex set V={v1,v2,,vn2,u1,w1}. The path viu1w1vj is a path of length 3 between vi,vj. The paths u1viw1vj and w1viu1vj are paths of length 3 between u1,vj and between w1,vj respectively. On the other hand, there exists no path of length 3 from u1 and w1. Now we can assign same color to u1, w1 and n2 other different colors to the vis. Thus χei(G)=n1.

3. If r4, and vi,wj are two vertices of two partite sets, then taking two vertices from two other partite sets xm,yl, one can construct a path vi,xm,yl,wj of length 3.

Let r=3 and G=Kk,l,m where two of k,l,m are at least 2. If k=1 and l,m2 with V(G)={v1,u1,ul,w1,,wm}, then v1uiwsuj, v1wiuswj, uiv1ujws, uiwtv1uj, wiuxv1wj with ij, give us a path of length 3 between v1,uj, v1,wj, ui,ws, ui,uj and wi,wj respectively.

Let r=3 and G=Kk,l,m where k,l,m2. Then, similar to the second part of situation 3, there exist a path of length 3 between any two vertices of G. Therefore χei(G)=n. Thus the result holds. ◻

Proposition 4.9. For fan graph Fm,n, we have

χei(Fm,n)={1,  m=1,n=2,m+1,  m2,n=2,m+n,  m=1,n4  and  m2,n3.

Proof. There are three situations to be considered.

1. If m=1,n=2. It is obvious.

2. If m2,n=2 and V(Fm,2)={v1,v2,,vm,u1,u2}, then by definition F2,2F1,3 and Fm,2=K¯mP2. Two vertices u1,u2 receive same color because there is no path of length 3 between them. On the other hand there exist a path viu1u2vj of length 3 between any pair of vertices vi,vj, and there exist a path viulvjuk with lk of length 3 between any pair of vertices vi,uk. Therefore χei(Fm,2)=m+1.

3.1. Let m=1,n4 and V(F1,n)={v1,u1,u2,,un}. Then v1ui+2ui+1ui, (in2), v1ui2ui1ui, (in1), uiv1uj+1uj, (i<j<n), uiv1un1un (i<n1) and un1un2v1un are the paths of length 3 between any pair of vertices ui, uj and v1, ui. Thus χei(F1,n)=1+n.

3.2. Let m2,n3. Using the reasons given in proof of part 3.1, one can easy to see that there is a path of length 3 between any pair of vertices of Fm,n.

All in all the proof is completed. ◻

5. Discussion and conclusions

From Propositions 2.1, 2.2 and 2.3,

1. Characterize graphs G with χei(G)=χ(G); Characterize graphs G with χei(G)=χi(G); Characterize graphs G with χei(G)=χ2(G).

2. After discussion on the solution of 1, one can revisit the Conjectures 1.1 and 1.3.
From Propositions 3.3 and 3.2, we can have the following.

3. Characterize graphs G with χei(G)=ρ(G).

4. Characterize graphs G with χei(G)=Δ(G)(Δ(G)1)2+1.

References:

  1. Y. Bu, A. Chen, D. Raspaud, and W. Wang. Injective coloring of planar graphs. Discrete Applied Mathematics, 157(4):663–672, 2009. https://doi.org/10.1016/j.dam.2008.08.016.
  2. Y. Bu, Z. Zhang, and H. Zhu. 2-distance coloring of planar graphs without triangles and intersecting 4-cycles. Discrete Mathematics, Algorithms and Applications, 15(02):2250084, 2023. https://doi.org/10.1142/S1793830922500847.
  3. Y. H. Bu and J. J. Cao. 2-distance coloring of planar graph. Discrete Mathematics, Algorithms and Applications, 13(02):2150007, 2021. https://doi.org/10.1142/S1793830921500075.
  4. G. Chartrand and P. Zhang. Chromatic Graph Theory. Chapman, Hall CRC Taylor, and Francis Group LLC, 1st edition, 2009.
  5. G. Ghazi, F. Rahbarnia, and M. Tavakoli. 2-distance chromatic number of some graph products. Discrete Mathematics, Algorithms and Applications, 12(02):2050021, 2020. https://doi.org/10.1142/S1793830920500214.
  6. G. Hahn, J. Kratochvíl, J. Širáň, and D. Sotteau. On the injective chromatic number of graphs. Discrete Mathematics, 256(1-2):179–192, 2002. https://doi.org/10.1016/S0012-365X(01)00466-6.
  7. J. Heuvel and S. McGuinness. Coloring the square of a planar graph. Journal of Graph Theory, 42:110–124, 2002. https://doi.org/10.1002/jgt.10077.
  8. X. Hu and B. M. Legass. Injective edge chromatic index of generalized Petersen graph P(ck, k). Discrete Mathematics, Algorithms and Applications, 16(01):2250189, 2024. https://doi.org/10.1142/S1793830922501890.
  9. F. Kramer and H. Kramer. A survey on the distance-colouring of graphs. Discrete Mathematics, 308(2-3):422–426, 2008. https://doi.org/10.1016/j.disc.2006.11.059.
  10. H. La and P. Valicov. Computer assisted discharging procedure on planar graphs: Application to 2-distance coloring. arXiv:2202.03885, 2022. https://doi.org/10.48550/arXiv.2202.03885.
  11. B. Lužar and R. Škrekovski. Counterexamples to a conjecture on injective colorings. Ars Mathematica Contemporanea, 8(2):291–295, 2015. https://doi.org/10.26493/1855-3974.516.ada.
  12. B. Lužar, R. Škrekovski, and M. Tancer. Injective colorings of planar graphs with few colors. Discrete Mathematics, 309(18):5636–5649, 2009. https://doi.org/10.1016/j.disc.2008.04.005.
  13. D. Mojdeh and B. Samadi. Further results on 2-distance coloring of graphs. Journal of Combinatorial Optimization, 45(1):1–12, 2023. https://doi.org/10.1007/s10878-022-00942-2.
  14. M. Molloy and M. R. Salavatipour. A bound on the chromatic number of the square of a planar graph. Journal of Combinatorial Theory, Series B, 94(2):189–213, 2005. https://doi.org/10.1016/j.jctb.2004.12.005.
  15. B. S. Panda. Injective coloring of some subclasses of bipartite graphs and chordal graphs. Discrete Applied Mathematics, 291(1):68–87, 2021. https://doi.org/10.1016/j.dam.2020.12.006.
  16. M. R. Raksha, P. Hithavarshini, C. Dominic, and N. K. Sudev. Injective coloring of complementary prism and generalized complementary prism graphs. Discrete Mathematics, Algorithms and Applications, 12(02):2050026, 2020. https://doi.org/10.1142/S1793830920500263.
  17. T. P. Sandhiya, J. Geetha, and K. Somasundaram. Total colorings of certain classes of lexicographic product graphs. Discrete Mathematics, Algorithms and Applications, 14(03):2150129, 2022. https://doi.org/10.1142/S1793830921501299.
  18. J. Song and J. Yue. Injective coloring of some graph operations. Applied Mathematics and Computation, 264(1):279–283, 2015. https://doi.org/10.1016/j.amc.2015.03.124.
  19. G. Wegner. Graphs with given diameter and a coloring problem. Technical report, University of Dortmund, Germany, 1–11, 1977. http://dx.doi.org/10.17877/DE290R-16496.
  20. D. B. West. Introduction to Graph Theory. Upper Saddle River: Prentice Hall, 7th edition, 2001.