Odd Colorings of Sparse Graphs
Abstract
A proper coloring of a graph is called odd if every non-isolated vertex has some color that appears an odd number of times on its neighborhood. The smallest number of colors that admits an odd coloring of a graph is denoted . This notion was introduced by Petruševski and Škrekovski, who proved that if is planar then ; they also conjectured that . For a positive real number , we consider the maximum value of over all graphs with maximum average degree less than ; we denote this value by . We note that is undefined for all . In contrast, for each , we give a (nearly sharp) upper bound on . Finally, we prove and . Both of these results are sharp.
1 Introduction
A proper coloring of a graph is odd††margin: odd coloring if every non-isolated vertex has some color that appears an odd number of times on its neighborhood. The smallest number of colors that admits an odd coloring of a graph is denoted ††margin: . Clearly, , since we can simply color each vertex with its own color. This notion was introduced by Petruševski and Škrekovski [13], who proved that if is planar then ; they also conjectured that .
Odd coloring is motivated by various types of hypergraph coloring. A hypergraph consists of a set of vertices and a set of (hyper)edges, each of which consists of an arbitrary set of vertices in . Most varieties of hypergraph coloring assign colors (integers in ) to the elements of subject to certain constraints. Standard hypergraph coloring requires only that no edge in is monochromatic. Even et al. [6] introduced conflict-free coloring, which requires that each edge in has some color that appears exactly once on its vertices. This topic has been widely studied [5, 8, 9, 12, 14]. Cheilaris et al. [4] studied odd coloring of hypergraphs, which requires that each edge in has some color that appears an odd number of times on its vertices. Aspects of this problem have been studied in [1, 2, 7]. For graphs, Cheilaris [3] studied conflict-free colorings with respect to open neighborhoods. That is, for each vertex some color appears exactly once on . Finally, Petruševski and Škrekovski [13] studied odd colorings of graphs, which are proper colorings where each vertex has some color that appears an odd number of times on . It is this parameter that we consider in the present short note.
The average degree of a graph is . The maximum average degree of a graph , denoted ††margin: , is the maximum, over all non-empty subgraphs of , of the average degree of . That is, . For each positive real number , let denote the family of graphs with . We denote by ††margin: , the maximum value of over all . The focus of this paper is bounding for various values of . We observe that is undefined for all . That is, there exists a sequence of graphs such that and for all . In contrast, for each , we give a (nearly sharp) upper bound on . We have two main results.
Theorem 1.
Fix such that . If , then . As , infinitely often there exists such that and .
When , we prove sharper upper bounds on .
Theorem 2.
Fix a graph . (a) If , then ; and (b) if , then . Furthermore, neither of these upper bounds on can be decreased, and neither of the inequalities can be weakened to allow equality.
For a proper coloring of some subgraph of , let ††margin: denote the unique color that appears an odd number of times in if such a color exists; otherwise, is undefined. Most of the rest of our notation and definitions are standard. But for the reader’s convenience we highlight a few terms. The girth††margin: girth of a graph is the length of a shortest cycle in (the girth of an acyclic graph is infinite). A -vertex††margin: //-vertex is a vertex of degree . A -vertex (resp. -vertex) is a vertex of degree at least (resp. at most) . A -neighbor††margin: //-neighbor of a vertex is a neighbor of that is a -vertex. Both -neighbor and -neighbor are defined analogously.
To close this introduction, we prove three easy results about graphs with .
Proposition 1.
A graph has if and only if has vertices but no edges. And has if and only if is bipartite and the degree of each vertex in is either 0 or odd.
Proof.
The first statement is obvious, so now we prove the second. If is bipartite, then has a proper 2-coloring . If the degree of each vertex is either 0 or odd, then is also an odd 2-coloring of . If is not bipartite, then has no proper 2-coloring, so clearly . Suppose instead is bipartite, but some vertex has positive even degree. Now the component of containing has only a single 2-coloring (up to permuting colors), but this 2-coloring is not odd since has only a single color, which appears an even number of times. ∎
Proposition 2.
If , then . If , then . Otherwise, .
Proof.
If , then we can repeat the colors around to get an odd 3-coloring. Suppose instead that . By the Pigeonhole principle, in every proper 3-coloring some color appears twice on the neighborhood of some vertex , so is not odd. Thus, . If and , then we can begin with or and continue with . Finally, it is easy to check that each proper 4-coloring of uses the same color on both neighbors of some vertex, so is not an odd coloring. Thus, . ∎
It is interesting to note that the class of graphs with and is richer than simply cycles with . Denote the vertices of by . Starting from such a graph, we can add arbitrarily many leaves at each with . It is straightforward to check that the resulting graph also has and . More generally, we can identify these with vertices of an arbitrary graph and will not decrease.
Proposition 3.
Every tree has . Thus, if , then . This bound on is sharp.
Proof.
Let be a tree. We use induction on . The case is trivial. Now suppose that . Let be a leaf of . Let and note that is a tree. By hypothesis, has an odd 3-coloring . Let be the neighbor of in . To extend to an odd 3-coloring of , we color with a color outside . This gives the desired odd 3-coloring of .
If , then is a forest. So the second statement follows from the first. Finally, the third statement follows from Proposition 2. ∎
The following proposition is folklore. But we include a proof for completeness.
Proposition 4.
If is a planar graph with girth at least , then .
Proof.
Fix a plane embedding of a planar graph with girth at least . Since each subgraph of also has girth at least , it suffices to show that . Summing the lengths of all facial walks gives , where is the set of all faces. To prove the proposition, we substitute this inequality into Euler’s formula and solve for . ∎
2
Recall, for each , that denotes the set of all graphs with ; and we write to denote the maximum value of over all . Proposition 3 shows that . In this section, we determine the set of all values such that is defined; when it is defined, we prove a (nearly sharp) upper bound on its value.
We denote by ††margin: the graph formed from by subdiving each edge once.
Lemma 5.
We have and .
Proof.
For each , let . To show that , we give a fractional orientation of where each vertex has indegree exactly . We orient each edge with fraction towards its endpoint of degree 2 and fraction toward its endpoint of degree . Each 2-vertex has indegree . Each -vertex has indegree . Thus, , as desired.
In an odd coloring of , all -vertices must get distinct colors. So . Given any coloring where all -vertices get distinct colors, it is easy to extend to an odd -coloring of . Thus, as claimed, . ∎
Corollary 6.
There exists a sequence such that for all and and for each there exists a graph with and .
Proof.
Let and . Now . ∎
Corollary 7.
is undefined whenever .
In Lemma 5, we considered which is formed by subdividing each edge of . Applying the same construction to any -chromatic graph yields a graph with and . Since there exist graphs with both chromatic number and girth arbitrarily large, there also exist graphs with and girth arbitrarily large, and with . However, every -chromatic graph that does not contain as a subgraph gives an upper bound on that is worse (larger) than that in Corollary 6. Consider an -critical (sub)graph with vertices and edges. Recall that . But is not -regular, by Brooks’ Theorem. Thus, . Subdividing each edge of gives with and . Thus, .
We next show that the construction in Lemma 5 is nearly sharp.
Theorem 1.
Fix such that . If , then . As , infinitely often there exists such that and .
Proof.
The second statement follows directly from Corollary 6.
For the first statement, let ††margin: and note that . Our proof is by induction on . The base case is when , so . Instead assume . If contains a 1-vertex with neighbor , then has an odd -coloring and we can extend to by coloring to avoid and . If contains a 3-vertex , then denote by . Now has an odd -coloring and we can extend to by coloring to avoid . So we assume that has neither 1-vertices nor 3-vertices. Suppose that has adjacent 2-vertices and , and denote the remaining neighbors of and , respectively, by and . Now has an odd -coloring . To extend this to and , we first color to avoid and then color to avoid the new color on as well as .
Since , we know that . We use discharging to show that some vertex with has “many” 2-neighbors. Similar to the case of adjacent 2-vertices above, we will delete and its 2-neighbors, find an odd -coloring for this smaller graph, and extend to all of . Let ††margin: . We give each vertex initial charge and use the following single discharging rule: Each 2-vertex takes charge from each neighbor. Since does not have adjacent 2-vertices, each 2-vertex finishes with charge . For each -vertex , let ††margin: denote the number of 2-neighbors of . Now finishes with charge . If for each , then we contradict the assumption . So there exists such that .
Form from by deleting and all of its 2-neighbors. By induction, has an odd -coloring . To extend to , we first color to avoid the colors on the colored neighbors (in ) of its deleted 2-neighbors (in ) as well as, for each -neighbor , to avoid and . For convenience, we denote this color used on by . Then we color each 2-neighbor of , with other neighbor , to avoid . This gives an odd coloring of . We must only show that uses at most colors. In total, must avoid at most colors. So it will suffice to show that . Note that . Since is an integer, . So now we must bound . Clearly, . So . Since is an integer, . Thus, . ∎
Remark 8.
In the previous proof we handled adjacent 2-vertices as follows. “Suppose that has adjacent 2-vertices and , and denote the remaining neighbors of and , respectively, by and . Now has an odd -coloring . To extend this to and , we first color to avoid and then color to avoid the new color on as well as .” In this argument we implicitly assumed that . Otherwise, our choice for color on might create a problem for . Specifically, suppose that is adjacent to 2-vertices and , which are also adjacent to each other, and that colors 1 and 2 each appear once on under (and no other color appears an odd number of times); this is possible when . Now is undefined. But if we are careless and color with 1 and color with 2, then we create a problem for . This obstacle has an easy work-around. Rather than defining relative to the coloring , we define it relative to the current coloring, including any vertices that were deleted but are now already colored. For brevity, we omit mention of this issue throughout the paper, since the solution above always works.
It is natural to consider a list-coloring analogue of odd coloring (although we omit a formal definition). We note that, with minor modifications, the proof of Theorem 1 works equally well in the context of odd list-coloring and even in the context of odd correspondence coloring. The same is true for the proofs of Theorem 2(a) and Theorem 2(b).
We suspect that the construction in Lemma 5 is sharp.
Conjecture 1.
Fix such that . If , then .
Conjecture 1 cannot be extended to allow , since , but , for each positive integer . In contrast, when and , we prove the conjecture in a stronger form. We show that if , then unless contains as a subgraph , where . We present these proofs in Sections 3 and 4.
Recall that a graph is -critical††margin: -critical if and for every edge . Analogously, a graph is odd--critical††margin: odd--critical if and for every edge . The latter notion is more subtle111An interesting example is , since , but has no odd-4-critical subgraph., since possibly is a subgraph of , but . More concretely, if we form from with by adding a leaf adjacent to each vertex, then , even though . It is easy to check that subdividing every edge of a -critical graph, with , yields an odd--critical graph. Kostochka and Yancey [10, 11] showed that the sparsest -critical graphs are so-called -Ore graphs. For each , by subdividing all edges in -Ore graphs, we get an infinite family of odd--critical graphs with maximum average degrees slightly larger than . We suspect that in the limit no infinite family of odd--critical graphs has smaller maximum average degree.
3 Odd 6-Colorings
The goal of this section is to prove the following result.
Theorem 2(a).
If , then . This includes all planar of girth at least 6.
Note that Theorem Theorem 2(a) is sharp in two senses. First, and , so the upper bound of 6 cannot be decreased. Second, and , so the hypothesis cannot be weakened at all.
Proof.
The second statement follows from the first by Proposition 4. So we prove the first. Assume the statement is false, and is a counterexample with as few vertices as possible. As in the proof of Theorem 1, we assume has no 1-vertex and has no adjacent 2-vertices. Suppose has a 3-vertex with neighbors , where . Denote the other neighbor of by . Let . By minimality, has an odd 6-coloring . To extend to , we color to avoid and then color to avoid the new color on and also to avoid . Thus, no 3-vertex has a 2-neighbor.
Suppose has a 4-vertex with neighbors , where . For each , denote the other neighbor of by . Let . By minimality, has an odd 6-coloring . To extend to , color to avoid , and then color each to avoid the new color on and also avoid , . Thus, no 4-vertex in has three (or more) 2-neighbors.
Suppose has a 5-vertex with five 2-neighbors . For each , denote the other neighbor of by . Let . By minimality, has an odd 6-coloring . To extend to , we color to avoid . We then color each to avoid the new color on , as well as . Thus, no 5-vertex in has five 2-neighbors.
Now we use discharging to reach a contradiction, which will finish the proof. We give each vertex initial charge and use the following single discharging rule: Each 2-vertex takes from each neighbor. We show that each vertex finishes with charge at least 3, which contradicts the hypothesis . Recall that .
If , then finishes with . If , then has no 2-neighbors, so finishes with 3. If , then has at most two 2-neighbors, so finishes with at least . If , then has at most four 2-neighbors, so finishes with at least . If , then finishes with at least , since . This finishes the proof. ∎
With a bit more analysis, we can prove the following stronger result.
Corollary 9.
If and does not contain as a subgraph, then .
Proof.
In the proof of Theorem Theorem 2(a), to contradict the hypothesis , we showed that each vertex finished with charge at least 3. To contradict the weaker hypothesis , it suffices to show that also some vertex finishes with charge more than 3. Suppose the contrary.
Now has no -vertex and each 6-vertex has six 2-neighbors. Each 5-vertex has exactly four 2-neighbors and each 4-vertex has exactly two 2-neighbors. Now it is straightforward to check that cannot have any two adjacent -vertices and ; in each case, we delete , , and their 2-neighbors, color the smaller graph and extend. Thus, is bipartite, where one part consists of 2-vertices and the other consists of 6-vertices. Form from by contracting one edge incident to each 2-vertex. Now is 6-regular, but no component is , so by Brooks’ Theorem. A 6-coloring of induces a 6-coloring of the 6-vertices in , which we can easily extend to an odd 6-coloring of . ∎
4 Odd 5-Colorings
The goal of this section is to prove the following result.
Theorem 2(b).
If , then . This includes all planar of girth at least 7.
Proof.
The second statement follows from the first by Proposition 4. So we prove the first. Assume the statement is false, and is a counterexample with as few vertices as possible. As in the proof of Theorem 1, we assume has no 1-vertex and has no adjacent 2-vertices. Suppose has a 3-vertex with neighbors , where . Denote the other neighbors of and by, respectively, and . Let . By minimality, has an odd 5-coloring . To extend to , we color to avoid and color each to avoid the new color on and avoid . Thus, each 3-vertex has at most one 2-neighbor. More generally, each 3-vertex has at most one -neighbor. If not, then we delete and any 2-neighbors. We color to avoid the color on each 3-neighbor, the color on the colored neighbor of each (deleted) 2-neighbor, as well as , where is the third neighbor of . Thus, each 3-vertex has at most one -neighbor.
Suppose has a 4-vertex with four 2-neighbors . For each , denote the other neighbor of by . Let . By minimality, has an odd 5-coloring . To extend to , color to avoid , and then color each , in succession, to avoid the new color on and also avoid , . Thus, no 4-vertex in has four 2-neighbors. Similarly, if each has degree at most 3 and at least one is a 2-vertex, then we form from by deleting and all its 2-neighbors. Again has an odd 5-coloring . We extend to by avoiding the color on each 3-neighbor and the color on the other neighbor of each 2-neighbor. Finally, we can extend the coloring to all 2-neighbors to get an odd 5-coloring of . Thus, has no 4-vertex with at least one 2-neighbor and all -neighbors. Similarly, does not contain adjacent 4-vertices, and , each with three 2-neighbors. If so, then has an odd 5-coloring. We color and with distinct colors that each differ from the colors on the three colored vertices at distance two in . Again, we can extend this coloring to an odd 5-coloring of .
Now we use discharging to reach a contradiction, which will finish the proof. We give each vertex initial charge and use the following two discharging rules.
-
(R1)
Each 2-vertex takes from each neighbor.
-
(R2)
Each 3-vertex with a 2-neighbor and each 4-vertex with three 2-neighbors takes from each -neighbor.
We show that each vertex finishes with charge at least 20/7, which contradicts the hypothesis . Recall that . If , then finishes with . If , then has at most one 2-neighbor. Further, does not give away charge by (R2), since no 3-vertex has both a 2-neighbor and a 3-neighbor, and also no 4-vertex has both three 2-neighbors and a 3-neighbor. So, if has a 2-neighbor, then it has two -neighbors and receives charge from each. Thus, finishes with at least . If has no 2-neighbor, then starts and finishes with 3.
Let be a 4-vertex, and recall that has at most three 2-neighbors. If has at most two 2-neighbors, then finishes with at least . If has three 2-neighbors, then its fourth neighbor is a -neighbor that does not receive charge from by (R2) but rather gives charge by (R2). So finishes with .
If is a -vertex, then finishes with at least , since . This finishes the proof. ∎
With a bit more analysis, we can prove the following stronger result. The proof is similar to that of Corollary 9, so we just provide a sketch.
Corollary 10.
If and does not contain as a subgraph, then .
Proof Sketch..
Assume is a counterexample. So has no -vertices and each 5-vertex in has five 2-neighbors. Each 3-vertex has a single 2-neighbor. Each 4-vertex either has exactly three 2-neighbors or has exactly two 2-neighbors and gives charge 1/7 to each of it -neighbors. In this case, we delete , its 4-neighbors and the 2-neighbors of all deleted 4-vertices. This smaller graph has an odd 5-coloring , and it is straightforward to check that we can extend to . So has no 4-vertices. This implies, in turn, that has no 3-vertices, since has no 3-vertex with two -neighbors. So is bipartite with vertices in one part of degree 2 and those in the other part of degree 5. We contract one edge incident to each 2-vertex, and 5-color the resulting graph by Brooks’ Theorem. Finally, we extend this 5-coloring to an odd 5-coloring of . ∎
References
- [1] D. P. Bunde, K. Milans, D. B. West, and H. Wu. Parity and strong parity edge-coloring of graphs. In Proceedings of the Thirty-Eighth Southeastern International Conference on Combinatorics, Graph Theory and Computing, volume 187, pages 193–213, 2007.
- [2] D. P. Bunde, K. Milans, D. B. West, and H. Wu. Optimal strong parity edge-coloring of complete graphs. Combinatorica, 28(6):625–632, 2008.
- [3] P. Cheilaris. Conflict-Free Coloring. 2009. Thesis (Ph.D.)–City University of New York.
- [4] P. Cheilaris, B. Keszegh, and D. Pálvölgyi. Unique-maximum and conflict-free coloring for hypergraphs and tree graphs. SIAM J. Discrete Math., 27(4):1775–1787, 2013.
- [5] P. Cheilaris and G. Tóth. Graph unique-maximum and conflict-free colorings. J. Discrete Algorithms, 9(3):241–251, 2011.
- [6] G. Even, Z. Lotker, D. Ron, and S. Smorodinsky. Conflict-free colorings of simple geometric regions with applications to frequency assignment in cellular networks. SIAM J. Comput., 33(1):94–136, 2003.
- [7] I. Fabrici and F. Göring. Unique-maximum coloring of plane graphs. Discuss. Math. Graph Theory, 36(1):95–102, 2016.
- [8] R. Glebov, T. Szabó, and G. Tardos. Conflict-free colouring of graphs. Combin. Probab. Comput., 23(3):434–448, 2014.
- [9] A. Kostochka, M. Kumbhat, and T. Łuczak. Conflict-free colourings of uniform hypergraphs with few edges. Combin. Probab. Comput., 21(4):611–622, 2012.
- [10] A. Kostochka and M. Yancey. Ore’s conjecture on color-critical graphs is almost true. J. Combin. Theory Ser. B, 109:73–101, 2014.
- [11] A. Kostochka and M. Yancey. A Brooks-type result for sparse critical graphs. Combinatorica, 38(4):887–934, 2018.
- [12] J. Pach and G. Tardos. Conflict-free colourings of graphs and hypergraphs. Combin. Probab. Comput., 18(5):819–834, 2009.
- [13] M. Petruševski and R. Škrekovski. Colorings with neighborhood parity condition. Dec. 2021, arXiv:2112.13710.
- [14] S. Smorodinsky. Conflict-free coloring and its applications. In Geometry---intuitive, discrete, and convex, volume 24 of Bolyai Soc. Math. Stud., pages 331--389. János Bolyai Math. Soc., Budapest, 2013.