Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Star Edge-Coloring of Square Grids

Star Edge-Coloring of Square Grids A star edge-coloring of a graph G is a proper edge-coloring without bichromatic paths or cycles of length four. The smallest integer k such that G admits a star edge-coloring with k colors is the star chromatic index of G. In the seminal paper on the topic, Dvořák, Mohar, and Šámal asked if the star chromatic index of complete graphs is linear in the number of vertices and gave an almost linear upper bound. Their question remains open, and consequently, to better understand the behavior of the star chromatic index, this parameter has been studied for a number of other classes of graphs. In this paper, we consider star edge-colorings of square grids; namely, the Cartesian products of paths and cycles and the Cartesian products of two cycles. We improve previously established bounds and, as a main contribution, we prove that the star chromatic index of graphs in both classes is either 6 or 7 except for prisms. Additionally, we give a number of exact values for many considered graphs. Keywords: star edge-coloring, star chromatic index, square grid, Cartesian product. 1 Introduction A proper edge-coloring of a graph G is called a star edge-coloring if there is neither bichromatic path nor bichromatic cycle of length four. The minimum number of colors for which G admits a star edge-coloring is called the star chromatic index and we denote it by  (G). st The star edge-coloring was defined in 2008 by Liu and Deng [8], and was motivated by the vertex version introduced by Grünbaum [4]. Despite a number of papers have already been published about this coloring, we have a very limited knowledge about it. In particular, the exact value of the star chromatic index of complete graphs is still not known, although some relatively strong lower and upper bounds have been determined by Dvořák et al. in their seminal paper [3]. Theorem 1 (Dvořák, Mohar, Šámal, 2013). The star chromatic index of the complete graph K satisfies p p 2 2(1+o(1)) log n 2n(1 + o(1))   (K )  n : st 1=4 (log n) 0 1+ In particular, for every  > 0 there exists a constant C such that  (K )  C n for every st n  1. Faculty of Applied Sciences, University of West Bohemia, Pilsen, Czech Republic. E-Mails: holubpre@kma.zcu.cz, mmockov@ntis.zcu.cz Faculty of Information Studies, Novo mesto, Slovenia. E-Mail: borut.luzar@gmail.com Faculty of Science, Pavol Jozef Šafárik University, Košice, Slovakia. E-Mails: erika.mihalikova@student.upjs.sk, roman.sotak@upjs.sk arXiv:2005.02864v2 [math.CO] 19 Aug 2020 They proved the upper bound using a nontrivial result about sets without arithmetic progressions, and up till now, it is still the best known. For the lower bound, they used an elegant double counting approach. The authors of [1] observed a small improvement in their proof and obtained the bound  (K )  3n(n 1)=(n + 4) (see [10] for a proof), which gives the exact values for st the chromatic index of K , for n 2 f1; 2; 3; 4; 8g. However, despite all efforts, the asymptotic behavior of the star chromatic index of complete graphs is not known, and in [3] the following question has been asked. Question 1 (Dvořák, Mohar, Šámal, 2013). What is the true order of magnitude of  (K )? st Is  (K ) = O(n)? st Another class of graphs with highly regular structure are complete bipartite graphs. They are important for better understanding of the coloring already on their own, and also, as Dvořák et al. [3] observed, the bounds for their star chromatic index provide bounds for the index of complete graphs. dlog ne 0 0 i2 0 (K ) n   (K )  2  (K i i ) : n;n n dn=2 e;dn=2 e st st st i=1 Recently, Casselgren et al. [2] considered complete bipartite graphs and proved the tight upper bound for K ; r  5, derived a lower and upper bound for K ; s  4, and, using computer, 3;r 4;s they also determined the star chromatic index for some complete bipartite graphs of small order. Star edge-coloring has been studied also for other classes of graphs, e.g., graphs with max- imum degree 3 [3, 6, 7] and 4 [14], subcubic Halin graphs [2], outerplanar graphs [1, 13], and planar graphs with various constraints [13]. Moreover, the list version of the star edge-coloring has also been investigated (see, e.g., [5, 9]). Finally, there is also a complexity result on the topic; namely, it is NP-complete to decide whether 3 colors suffice for a star edge-coloring of a subcubic multigraph [6]. Since most of the obtained upper bounds for the star chromatic index are not tight and many questions remain open, we focus our attention to graphs with a relatively simple structure, i.e. to the Cartesian products of graphs. The star edge-coloring of the Cartesian products of graphs has already been considered by Omoomi and Dastjerdi [11]. They established an upper bound for the star chromatic index of the Cartesian product of two arbitrary graphs, proved its exact values for the Cartesian product of two paths (Theorem 4), and they started investigation on the Cartesian products of a path and a cycle, and the Cartesian product of two cycles (i.e., square grids). They further proved upper bounds for d-dimensional grids and d-dimensional hypercubes. Motivated by the results presented in [11], in this paper, we consider star edge-coloring of square grids; in particular, the Cartesian products of two cycles and the Cartesian products of paths and cycles. Apart from the usual combinatorial methods, due to the complexity of the problems considered in this paper, we have used computer to obtain star edge-colorings of small graphs and to establish some of the lower bounds. Standard (formal) mathematical proofs would require enourmous amount of case analysis, while their contribution to the theory would be minimal. We establish exact bounds for the star chromatic index of many graphs from the two considered classes, and show that the upper bound for the chromatic index of both Cartesian products is 7. The paper is structured as follows. We give our notation and prove some auxiliary results in Section 2. Section 3 contains the algorithm used in our computations and describes the 2 preprocessing procedures used in them. In Section 4, we present the main results of this paper, and we list some open problems in Section 5. 2 Preliminaries In this section, we present some additional terminology used in the paper and give auxiliary results. We abbreviate a ‘star edge-coloring with k colors’ to a ‘star k-edge-coloring’, and, if it is clear from the context, sometimes we just write ‘coloring’ instead of ‘star edge-coloring’. The Cartesian product of graphs G and H, denoted by G  H, is the graph with the vertex 0 0 set V (G) V (H ) and edges between the vertices (u; v) and (u ; v ) if: 0 0 uu 2 E(G) and v = v (a G-edge), or 0 0 u = u and vv 2 E(H ) (an H-edge). We call the graphs G and H the factor graphs. The G-fiber with respect to v 2 V (H ), denoted by G , is the copy of G in G  H induced by the vertices having v as the second component. Analogously, the H-fiber with respect to u 2 V (G), denoted by H , is the copy of H in G  H induced by the vertices having u as the first component. Since the Cartesian product of two paths is a subgraph of the Cartesian product of a path and a cycle, and the Cartesian product of a path and a cycle is a subgraph of the Cartesian products two cycles, we have the following sequence of inequalities. Observation 1. For every pair of positive integers m and n, where m  3 and n  3, we have 0 0 0 (P  P )   (C  P )   (C  C ): m n m n m n st st st Having a star edge-coloring of the Cartesian product of an n-cycle and a graph H, we can extend it to a coloring of the Cartesian product of a cycle of length k  n and H. Lemma 1. For every integers k and m, where k  2 and m  3, and for every graph H, we have 0 0 (C  H )   (C  H ) : km m st st Proof. Let C = u : : : u u , C = v : : : v v , and V (H ) = fw ; : : : ; w g. Moreover, let m 1 m 1 km 1 km 1 1 n p : f1; : : : ; kmg ! f1; : : : ; mg be an assignment given by p(t) = s if and only if (ts) is divisible by m. Let  be a star edge-coloring of C  H. Consider an edge e = (v ; w )(v ; w ) of C  H. m i a j b km By the definition of the Cartesian product, we have i = j or a = b. Note that if a = b, then ji jj = 1, and thus we may assume j = i + 1. We define a proper edge-coloring  of C  H km as follows. If i = j, then set  ((v ; w )(v ; w )) = ((u ; w )(u ; w )). In the case a = b, we i a i b a b p(i) p(i) set  ((v ; w )(v ; w )) = ((u ; w )(u ; w )). i a j a p(i) a p(j) a Now we show that  is also a star edge-coloring. For an integer s, where 1  s  k  m, let G be the graph induced by the vertices f(v ; w)g, where ` 2 fs + 1; : : : ; s + mg (the values s ` s + 1; : : : ; s + m are taken modulo k  m) and all w 2 V (H ), i.e., G is the graph induced on m consecutive H-fibers. Observe that the coloring  on G corresponds to a coloring  of the subgraph of C H without the edges (u ; w)(u ; w), for all w 2 V (H ). Therefore, p(s+m) p(s+1) every 4-path and every 4-cycle in C  H, contained in some G , is not bichromatic. km s Finally, if a 4-path or a 4-cycle is not contained in any G , then it contains at least m edges of type (v ; w )(v ; w ) (i.e., only when m 2 f3; 4g). However, in the case of m = 3, i a j a 3 three consecutive edges on every C -fiber receive three distinct colors, and hence no 4-path with three consecutive edges on a C -fiber is bichromatic. If a 4-path has two consecutive edges on a C -fiber, an edge in an H-fiber, and the fourth edge in another C -fiber, then its coloring m m corresponds to a coloring of some 4-path in C  H, which is not bichromatic. In the case of m = 4, we only have 4-cycles, whose colorings correspond to a coloring of a 4-cycle by  in some C -fiber, and hence they are not bichromatic. We continue by showing how star edge-colorings of two Cartesian products, each having at least one cycle as a factor, can be combined. Let m and n be a pair of integers, where 3  m < n, and let v ; : : : ; v be consecutive vertices of the cycle C . We say that a star edge-coloring  of 1 n n C  H includes a star edge-coloring of C  H if the coloring  of the subgraph of C  H n m n induced by the vertices of m consecutive H-fibers H ; : : : ; H , together with the additional v v 1 m edges e = (v ; w)(v ; w), for all w 2 V (H ), where we set  (e ) = ((v ; w)(v ; w)), is a star w 1 m w 1 n edge-coloring. Symmetrically, we can say that a star k-edge-coloring of H C includes a star edge-coloring of H  C . Note that the star edge-coloring of C  H, constructed in the proof of Lemma 1, m km includes a star edge-coloring of C  H. Lemma 2. If for a pair of positive integers m and n, where m < n, a star k-edge-coloring of C  H includes a star edge-coloring of C  H, then, for every pair of non-negative integers p n m and q, we have (C  H )  k : pm+qn st Proof. Let  be a star k-edge-coloring of C H which includes a star edge-coloring  of C H. n m Let C = v : : : v v , C = v : : : v v , and C = u : : : u u . Furthermore, we define n 1 n 1 m 1 m 1 pm+qn 1 pm+qn 1 an assignment r : f1; : : : ; pm +qng ! f1; : : : ; ng such that, if t  pm, then r(t)  m and tr(t) is divisible by m, and, if t > pm, then t pm r(t) is divisible by n. Now, similarly as in the proof of Lemma 1, we define an edge-coloring  of C  H. We pm+qn combine p copies of  followed by q copies of . More precisely, for w ; w 2 H, a b ((v ; w )(v ; w )); for i = j  pm ; > a r(i) r(i) b ((v ; w )(v ; w )); for i = j > pm ; r(i) a r(i) b ((u ; w )(u ; w )) = i a j b ((v ; w )(v ; w )); for i = j 1  pm ; > a a r(i) r(j) ((v ; w )(v ; w )); for i = j 1 > pm : a a r(i) r(j) Note that, by the definition of  , we have  ((v ; w)(v ; w)) = ((v ; w)(v ; w)). m 1 n 1 It remains to show that  is a star edge-coloring. For an integer s, where 1  s  pm + qn, let G be the graph induced by the vertices of m + 1 consecutive H-fibers H ; : : : ; H s u u s+1 s+m+1 (the indices s + i are taken modulo pm + qn). Since the coloring of each G is a part of two consecutive copies of  or a part of two consecutive copies of , no 4-path and no 4-cycle in G is bichromatic (using Lemma 1 for k = 2). Finally, if a 4-path is not contained in any G , then it contains 4 edges of type (v ; w )(v ; w ) s i a j a and m = 3. Moreover, such a 4-path traverses the H -fibers, for i 2 fpm; : : : ; pm + 4g or i 2 fpm + qn; 1; : : : ; 4g. In both cases, colors of three consecutive edges of the 4-path correspond to colors of a C -fiber of  and therefore they are distinct. This completes the proof. We will use Lemma 2 to prove results for arbitrary lengths of cycles. To do that, we will use the following result on Frobenious numbers [12]. 4 Theorem 2 (Sylvester, 1882). Let positive integers n and m be relatively prime. Then for every integer k  (n 1)(m 1) there exist non-negative integers and such that k =  n +  m : We also recall the result of Dvořák, Mohar, and Šámal [3] about star edge-coloring of subcubic graphs, which we will use when considering prisms. Theorem 3 (Dvořák, Mohar, Šámal, 2013). (a) If G is a subcubic graph, then  (G)  7. st (b) If G is a simple cubic graph, then  (G)  4, and the equality holds if and only if G covers st the graph of the 3-dimensional hypercube. Here, a graph G is said to cover a graph H if there is a graph homomorphism from G to H that is locally bijective. In other words, there is a mapping f : V (G) ! V (H ) such that whenever uv is an edge of G, the image f (u)f (v) is an edge of H, and, for each vertex v 2 V (G), f is a bijection between the neighbors of v and the neighbors of f (v). At this point, we remark the following, somehow hidden, corollary of the above result. Hexag- onal grids are subcubic graphs and they cover the graph of the 3-dimensional hypercube. Thus: Corollary 1. For an infinite hexagonal grid G, we have (G) = 4 : st 3 Computer Computations and Algorithm For our computations, we used a simple backtracking algorithm, which, together with some preprocessing, enabled us to compute exact lower bounds for some important cases on one hand, and on the other hand, provided star edge-colorings with required properties for some graphs. The main coloring algorithm takes three input parameters: the graph to be colored, the number of colors, and a possible precoloring of some edges, in order to avoid testing some isomorphic partial colorings; e.g., one may fix the colors on the edges incident to a vertex of maximum degree. Note also that before calling the function StarColor, we first verify that the precoloring of the edges is a star edge-coloring. Another important part of our algorithm is determining the order of edges (the function GetEdgeOrdering), in which it tries to color them. We order the edges (ignoring the precolored edges) by the number of precolored neighbors (incident edges) and the number of neighbors appearing earlier in the ordering in a descending order. The function Conflict checks if assigning a color to the current edge introduces a conflict, namely, it checks if two adjacent edges receive the same color, and if a bichromatic 4-path or 4-cycle appears. In some cases, we manually controlled the different cases of precolored edges. If we established some additional property of a required coloring, e.g., that no 4-path can be colored with just three colors, we included that in the procedure. Finally, we also adopted the algorithm to output all possible colorings of a given graph, and in the case of symmetric graphs, e.g., cycles, we eliminated isomorphic colorings. The remaining colorings were used to test if they can be extended to graphs on more vertices. 5 Algorithm 1 Star edge-coloring algorithm 1: procedure StarColor(G; k;P) . Graph G, number of colors k, precolored edges P 2: edgeOrder GetEdgeOrdering(G, P) 3: edgeColors InitEdgeColors(P,edgeOrder) 4: triedColors InitTriedColors(edgeOrder) . Dictionary of empty lists for all edges 5: for i in 1::edgeOrder.Count do . Try to color edges according to the ordering 6: e edgeOrder[i] 7: isColored false 8: for color c in f1::kg n triedColors[e] do 9: edgeColors[e] c 10: if Conflict(c, G, edgeColors) then . Check if a conflict occurs 11: edgeColors[e] ; 12: else 13: add c to triedColors[e] 14: isColored true 15: goto 18 16: end if 17: end for 18: if !isColored and i > 1 then . If no color is found, continue if not at first edge 19: edgeColors[edgeOrder[i 1]] ; 20: i = i 2 . Step up, 2 handles automatic loop increment 21: else if !isColored and i == 1 then 22: return “No coloring found” 23: else if isColored and i == edgeOrder.Count then 24: return edgeColors . A star edge-coloring is found 25: end if 26: end for 27: end procedure 4 Cartesian products of paths and cycles 4.1 Cartesian products of paths In a recent paper, Omoomi and Dastjerdi [11] established tight bounds for the star chromatic index of two paths (see Table 1). Theorem 4 (Omoomi and Dastjerdi, 2019). For the graph P P , where m and n are integers m n with 2  m  n, we have 3; if m = n = 2; 4; if m = 2 and n  3; (P  P ) = m n st 5; if m = 3 and 3  n  4; 6; otherwise. As a corollary, we establish the lower bound of 6 colors for the Cartesian products, where one factor is a cycle and the other is a path of length at least 2. 6 + mnn 2 3 4 5 2 3 4 4 4 3 4 5 5 6 4 4 5 6 6 5 4 6 6 6 Table 1: The star chromatic index of the Cartesian products of two paths (P  P ) m n st Corollary 2. For every pair of integers m and n, where m  3 and n  3, we have (C  P )  6: m n st Proof. We first note that the graph C P is one of the two known examples of simple bridgeless 3 2 cubic graphs that have star chromatic index equal to 6 [9]. Then, for m = 3, by Observation 1, 0 0 we have  (C  P )   (C  P ) = 6. 3 n 3 2 st st If m = 4, we proceed by a contradiction. Suppose that  (C  P )  5. Then, by Lemma 1, 4 n st 0 0 (C  P )  5, for any integer `, and hence also  (P  P )  5, a contradiction. Finally, n 5 n 4` st st 0 0 if m  5, then we have  (C  P )   (P  P ) = 6 by Theorem 4 and Observation 1. m n 5 3 st st 4.2 Cartesian products of cycles Having the Cartesian products of paths resolved, the logical direction of research is consideration of cylinders and toroidal grids, i.e., the Cartesian products of cycles and paths, and the Cartesian products of two cycles. We begin by giving some results about the latter. Corollary 2 implies that the Cartesian product of any two cycles will need at least 6 colors for a star edge-coloring. On the other hand, as we will show in this subsection, the star chromatic index of the Cartesian product of two cycles is at most 7. We first investigate the Cartesian products of C with another cycle. Theorem 5. For every integer n, where n  3, we have 6; if n = 3k, (C  C ) = 3 n st 7; otherwise Proof. By Corollary 2,  (C  C )  6. Now suppose that n = 3k for some integer k  1. If 3 n st k = 1, then there is a star 6-edge-coloring of C  C (one is depicted in Figure 1(a)). Next, by 3 3 Lemma 1, we have  (C  C ) = 6. 3 n st Using Algorithm 1, we infer that the Cartesian product C  P has only one star 6-edge- 3 3 coloring up to a permutation of colors. Namely, three colors, say 0, 1, and 2, appear on C -fibers, and the colors 4, 5, and 6 on the P -fibers. Since C  P is a subgraph of every graph C  C , 3 3 3 3 n it follows that such Cartesian products admit a star 6-edge-coloring only when n is divisible by Therefore, if n 6= 3k for every integer k, then  (C  C )  7. In Figures 1(d) and 1(e), 3 n st a star 7-edge-coloring of C  C and C  C , respectively, is depicted. Observe that, in both 3 7 3 8 colorings, a star 6-edge-coloring of C  C is included. Hence, by Lemma 2 and Theorem 2, we 3 3 have  (C  C ) = 7 for every n, n  7, not divisible by 3. The remaining two cases, namely 3 n st n = 4 and n = 5, are depicted in Figures 1(b) and 1(c), respectively. 7 Similarly as in the proof of Theorem 5, we can use Lemma 1 (twice) to extend the star edge-coloring of C  C to products of cycles of lengths divisible by 3. 3 3 Corollary 3. For every pair of positive integers k and `, we have (C  C ) = 6 : 3k 3` st We proceed with a result about the Cartesian products of C with another cycle. Theorem 6. For every pair of positive integers k and `, where k  1 and `  2, we have (C  C ) = 6 : 4k 2` st Proof. We use the star 6-edge-coloring  of C  C depicted in Figure 3. Note that 10 4 10 10 includes a star 6-edge-coloring C  C , and a star 6-edge-coloring C  C . Therefore, by 4 4 4 6 Lemma 2 and Theorem 2, we have  (C  C ) = 6 for every integer `  4. Finally, we use 4 2` st Lemma 1 to infer  (C  C ) = 6 for every integer k. 4k 2` st Proposition 1. For n 2 f5; 7; 9; 11g, we have (C  C ) = 7 : 4 n st Proof. Using Algorithm 1, we established that  (C  C ) > 6 for every n 2 f5; 7; 9g. The 4 n st bounds  (C  C ) = 7 follow from the star 7-edge-colorings depicted in Figures 2(a), 2(c), 4 n st and 2(d). In the case of n = 11, we split the computation in two steps. First, using Algorithm 1, we determined that if the edges of some C -fiber are colored in such a way that a same color appears twice on some 4-path, then the coloring cannot be extended to a star 6-edge-coloring of C  C . In the second step, the algorithm checked only the colorings in which every 4-path 4 11 in each C -fiber had four colors on its edges. It turned out that such a coloring does not exist. Therefore,  (C  C ) = 7 by the star edge-coloring depicted in Figure 2(c) and Lemma 1. 4 11 st Theorem 7. For any odd integer n, where n  13, we have (C  C )  7 : 4 n st Proof. In Figure 2(c), we present a star 7-edge-coloring of C C with a star 7-edge-coloring of 4 7 C  C included. Thus, by Lemma 2 and Theorem 2, we infer that  (C  C )  7 for every 4 4 4 n st odd n, where n > 18. Colorings for n 2 f13; 15; 17g can be obtained by using Lemma 2 and the colorings depicted in Figures 2(a) (for n = 15) and 2(d) (for n 2 f13; 17g). Theorem 8. For every integer n, where n  3, we have (C  C ) = 7 : 5 n st Proof. The lower bounds  (C  C ) > 6 for 3  n  6, were established using Algorithm 1. 5 n st For n  7, using Algorithm 1, we infer that  (C  P )  7. Therefore, by Observation 1, we 5 n st have  (C  C )  7. 5 n st Star 7-edge-colorings of C  C , for m 2 f3; 4; 5; 7; 11g, are depicted in Figures 1(c), 2(a), 5 m 4(a), 4(b), and 4(c), respectively. By Lemma 1, we also infer star 7-edge-colorings of C  C 5 m for m 2 f6; 8; 9; 10g. Finally, since in Figure 4(b), a star 7-edge-coloring of C  C is included, 5 3 by Lemma 2 and Theorem 2, we obtain  (C  C ) = 7 for every integer n  12. 5 n st 8 Theorem 9. For every integer n, where n  3, we have 6; if n  0 mod 3 or n  0 mod 4, (C  C ) = 6 n st 7; otherwise. Proof. By the star 6-edge-colorings depicted in Figures 1(a) and 2(b), and by Lemma 1, we have (C  C ) = 6, for every integer n divisible by 3 or 4. 6 n st Now we show that  (C  C ) > 6 if n is not divisible by 3 or 4. First, we consider the 6 n st graph C P , where P = v : : : v . We start with a precolored C -fiber at the vertex v (i.e., 6 31 31 1 31 6 16 the middle C -fiber) using each of the nine possible star 6-edge-colorings of C (up to symmetries 6 6 and permutations of colors). Using Algorithm 1, we tried to extend such a precoloring to the whole C  P . For five colorings of the C -fiber, namely for (0; 1; 0; 2; 0; 3), (0; 1; 0; 2; 1; 2), 6 31 6 (0; 1; 0; 2; 1; 3), (0; 1; 2; 0; 1; 3), and (0; 1; 2; 0; 3; 4), we obtain that such precolorings cannot be extended. For the remaining four precolorings, namely (0; 1; 0; 2; 3; 2), (0; 1; 0; 2; 3; 4), (0; 1; 2; 0; 1; 2), and (0; 1; 2; 3; 4; 5), we obtain 27 078 colorings of C  P in total. Some of them are either 4-, 6 31 or 6-periodical, i.e., the initial coloring repeats on every 4-th or 6-th fiber, except at the final three fibers on both sides, where the coloring restrictions are relaxed. The remaining 26 448 colorings correspond to the precoloring (0; 1; 2; 3; 4; 5), and moreover, all C -fibers are colored by shifts of this precoloring, and every pair of adjacent C -fibers is either 6 6 colored with the same sequence of colors, or the coloring of one is the coloring of the other shifted by 1. In a more detailed analysis of these colorings, we find that, if they are periodic, then the period must be a multiple of 6. Thus, for the graphs C  C , it follows that they are star 6-edge-colorable if n is divisible 6 n by 3 or 4. Otherwise they are not star 6-edge-colorable. Theorem 10. For every integer n, where n  3, we have (C  C ) = 7 : 7 n st Proof. The lower bounds  (C  C ) > 6, for n 2 f3; 4; 5; 6g, were established using Al- 7 n st gorithm 1. For n  7, using Algorithm 1, we infer that  (C  P )  7. Therefore, by 7 n st Observation 1, we have  (C  C )  7. 7 n st Star 7-edge-colorings of C  C , for n 2 f3; 4; 5; 7g, are depicted in Figures 1(d), 2(c), 4(b), 7 n and 5, respectively. By Lemma 1, from these colorings, we also infer star 7-edge-colorings of C  C for n 2 f6; 8; 9; 10g. Moreover, in the coloring depicted in Figure 6, a star 7-edge- 7 n coloring of C  C is included, and hence we also have  (C  C ) = 7. Finally, since in the 7 11 7 11 st coloring depicted in Figure 5, a star 7-edge-coloring of C  C is included, by Lemma 2 and 7 3 Theorem 2, we have  (C  C ) = 7 for every n  12. 7 n st Proposition 2. For C  C , we have 8 9 (C  C ) = 7 : 8 9 st Proof. We determined that  (C  C )  7 by exhaustive computer search. Namely, we gener- 8 9 st ated all 147 non-isomorphic star 6-edge-colorings of C and tried to extend each of them to the graph C  C . None of them could be extended, thus  (C  C )  7. The equality follows 8 9 8 9 st from Lemma 1 and the fact that  (C  C ) = 7. 4 9 st 9 Finally, we give a general result, showing that 7 is the upper bound for the star chromatic index of the Cartesian products of any two cycles. Theorem 11. For every pair of positive integers m and n, where 3  m  n, we have (C  C )  7 : m n st Proof. By Theorems 5-10, we have  (C  C )  7 for 3  m  7 and n  3. Furthermore, m n st by Lemma 1, we can use Theorem 7 to obtain a star 7-edge-coloring of C  C , Theorem 5 8 n to obtain a star 7-edge-coloring of C  C , and Theorem 8 to obtain a star 7-edge-coloring of 9 n C  C , for every n  3. The star 7-edge-coloring of C  C is depicted in Figure 6. 10 n 11 11 We complete the proof by showing that  (C  C )  7 if m; n  12. Note that the star m n st 7-edge-coloring of C C depicted in Figure 5, includes a 7-edge-coloring of C C and a 7-edge- 7 7 3 7 coloring of C C . Furthermore, the latter two colorings include a common star 7-edge-coloring 7 3 of C  C . This fact enables us to use Lemma 2 and Theorem 2 to obtain  (C  C )  7 3 3 m n st for m; n  12. The above results are summarized in Table 2. mnn 3 4 5 6 7 8 9 10 11 12 3 6 7 7 6 7 7 6 7 7 6 4 7 6 7 6 7 6 7 6 7 6 5 7 7 7 7 7 7 7 7 7 7 6 6 6 7 6 7 6 6 7 7 6 7 7 7 7 7 7 7 7 7 7 7 8 7 6 7 6 7 6 7 6 7 6 9 6 7 7 6 7 7 6 7 7 6 10 7 6 7 7 7 6 7 7 7 6 11 7 7 7 7 7 7 7 7 7 7 12 6 6 7 6 7 6 6 6 7 6 Table 2: The star chromatic index of the Cartesian products of cycles (C  C ). In red, we denote the cases, where the exact bounds are not estab- m n st lished yet. The value 7 means that the exact value of the star chromatic index is either 6 or 7. 4.3 Cartesian products of cycles and paths In the last part of this section, we give results about the Cartesian products of paths and cycles. We begin with proving the cases for specific lengths of cycles. Theorem 12. For every pair of integers k and n, where k  2 and n  3, we have (C  P ) = 6 : 2k n st Proof. By Corollary 2, we have  (C  P )  6. On the other hand, 6 colors also suffice by 2k n st Theorem 6, since C  P is a subgraph of C  C for every `  n=4. 2k 2k 4` 10 Theorem 13. For every pair of integers k and n, where n  3, we have (C  P ) = 6 : 3k n st Proof. By Corollary 2, we have  (C  P )  6. On the other hand, 6 colors also suffice by 3k n st Corollary 3, since C  P is a subgraph of C  C for every `  n=3. 3k n 3k 3` Theorem 14. For every integer n, where n  3, we have 6; if n 2 f3; 4; 5; 6g, (C  P ) = 5 n st 7; if n  7. Proof. Let G = C P for some integer n  3. Suppose first that n 2 f3; 4; 5; 6g. By Theorem 4, 5 n 0 0 we have  (P  P ) = 6, and thus, since P P is a subgraph of G, it follows that  (G)  6. 3 5 5 3 st st On the other hand, in Figure 7(a), we give a star 6-edge-coloring of C  P , hence establishing 5 6 (C  P ) = 6 for every n 2 f3; 4; 5; 6g. Now, suppose that n  7. Using Algorithm 1, we 5 n st 0 0 infer that  (C  P )  7. The upper bound  (C  P )  7 follows from the fact that 5 n 5 n st st (C  C ) = 7 for every positive integer k (see Theorem 8 and Figure 4(a)). 5 5k st Theorem 15. For every integer n, where n  3, we have 6; if n 2 f3; 4; 5; 6g, (C  P ) = 7 n st 7; if n  7. Proof. Let G = C  P for some integer n  3. Suppose first that n 2 f3; 4; 5; 6g. Since 7 n 0 0 (P  P ) = 6, we again have  (G)  6. On the other hand, in Figure 7(b), we give a 3 5 st st star 6-edge-coloring of G, and hence  (G) = 6 for every n 2 f3; 4; 5; 6g. Suppose now that st 0 0 n  7. Using Algorithm 1, we infer that  (C  P ) = 7, and hence  (G)  7. The equality 7 7 st st is established by Figure 1(d), where a pattern for a star 7-edge-coloring of C C is presented. 7 3k Since G is a subgraph of C  C , for k large enough, the statement follows. 7 3k We now turn our attention to the Cartesian products of cycles and paths P , for m 2 f2; 3; 4g. Theorem 16. For every integer m, where m  3, we have >6; if m = 3, (C  P ) = 4; if m  0 mod 4, m 2 st 5; otherwise. Proof. For m = 3, recall that  (C  P ) = 6. If m  0 mod 4, then the graph C  P covers 3 2 m 2 st the graph Q , and hence its star chromatic index equals 4 by Theorem 3. Finally, if m 6 0 mod 4, by Theorem 3, we have  (C  P )  5. The equality follows from Theorem 7 in [11]. m 2 st Theorem 17. For every pair of integers m and n, where m  3 and n 2 f3; 4; 5; 6g, we have (C  P ) = 6 : m n st 11 Proof. First, recall that 6 colors are needed in all the cases by Corollary 2. Next, since C  P m 6 contains all the graphs C  P , for n 2 f3; 4; 5g, it suffices to show that there exists a star m n 6-edge-coloring of C  P . m 6 0 0 By Lemma 1, Theorems 5 and 6, we have  (C  P ) =  (C  P ) = 6, for any positive 3k 6 4k 6 st st integer k. Similarly, the star 6-edge-colorings depicted in Figures 7(a) and 7(b) together with 0 0 Lemma 1 imply  (C  P ) =  (C  P ) = 6. Furthermore, using the star 6-edge-coloring 5k 6 7k 6 st st of C  P depicted in Figure 7(c), which includes a star 6-edge-coloring of C  P , we obtain 10 6 3 6 (C  P ) = 6 for every m 2 f10; 13; 16g. Moreover, together with Lemma 2 and Theorem 2, m 6 st this coloring implies  (C  P ) = 6 for every m  18. For the remaining two cases, namely m 6 st m 2 f11; 17g, the corresponding colorings are depicted in Figures 8(a) and 8(b) (note that, in fact, we have even more, namely, we give a star 6-edge-coloring of C  P and a star 6-edge- 11 8 coloring of C  P which includes a star 6-edge-coloring of C  P ). 14 8 3 8 We collect known and our new results in Table 3. mnn 2 3 4 5 6 7 8 9 3 6 6 6 6 6 6 6 6 4 4 6 6 6 6 6 6 6 5 5 6 6 6 6 7 7 7 6 5 6 6 6 6 6 6 6 7 5 6 6 6 6 7 7 7 8 4 6 6 6 6 6 6 6 9 5 6 6 6 6 6 6 6 10 5 6 6 6 6 6 6 6 11 5 6 6 6 6 6 6 7 12 4 6 6 6 6 6 6 6 13 5 6 6 6 6 7 7 7 14 5 6 6 6 6 6 6 6 15 5 6 6 6 6 6 6 6 16 4 6 6 6 6 6 6 6 17 5 6 6 6 6 6 6 7 18 5 6 6 6 6 6 6 6 19 and m  0 mod 4 4 6 6 6 6 6 6 6 19 and m  r mod 12, r 2 f2; 3; 6; 9; 10g 5 6 6 6 6 6 6 6 19 and m  r mod 12, r 2 f1; 5; 7; 11g 5 6 6 6 6 7 7 7 Table 3: The star chromatic index of the Cartesian products of cycles and paths (C  P ). In red, we denote the cases, where the exact bounds are not estab- m n st lished yet. The value 7 means that the exact value of the star chromatic index is either 6 or 7. 5 Conclusion In this paper, we have established tight upper bounds for the Cartesian product of cycles and paths, and the Cartesian products of two cycles. We proved that 7 colors always suffice for 12 star edge-colorings of these graphs, which is, in a way, a surprising bound, especially because at least 6 colors are always needed as soon as one of the factors is not isomorphic to P . Although we improved existing bounds and proved a number of exact values, there are still some open questions. We are very confident that the following conjecture is true. Conjecture 1. There exist constants K and K such that for every pair of integers m and n, 1 2 where m  K and n  K , we have 1 2 (C  P ) = 6 : m n st Note that Conjecture 1 is equivalent to the following. Conjecture 2. There exists a constant K such that for every integer m, where m  K, there exists an integer n such that (C  C ) = 6 : m n st In fact, we believe (with a bit lower confidence) that the following stronger version of Con- jecture 1 can also be confirmed. Conjecture 3. There exists a constant L such that for every pair of integers m and n, where m; n  L, we have (C  C ) = 6 : m n st The above conjectures seem to be challenging, since we were not able to observe any straight- forward pattern in colorings of different Cartesian products, despite the help of computer in our constructions. There are also some (maybe) less complicated open questions regarding the Cartesian product of the cycles C , m 2 f11; 13; 17g, with paths of arbitrary lengths. Question 2. What is the star chromatic index of C  P for m 2 f11; 13; 17g and n  7? m n As we observed in Theorems 14 and 15, the star chromatic index increases for the Cartesian products of the cycles C and C with paths on at least 7 vertices. We expect this phenomenon 5 7 will repeat also for some cycle C , where m 2 f11; 13; 17g, although, in the case of C and C , m 11 17 we found star 6-edge-colorings of C  P and C  P , which could indicate that 6 colors are 11 8 17 8 sufficient in these cases. To conclude, as in the case of complete graphs, also for the Cartesian products of paths and cycles (and two cycles), it is hard to determine their chromatic index with our current methods. However, it seems that the latter will be easier to resolve as the former. Acknowledgment. PH and MM acknowledge the financial support from the project GA20– 09525S of the Czech Science Foundation. BL was partly supported by the Slovenian Research Agency Program P1–0383 and the project J1–1692. EM and RS were supported by the Slovak Research and Development Agency under the Contract No. APVV–15–0116 and by the Slovak VEGA Grant 1/0368/16. References [1] L’. Bezegová, B. Lužar, M. Mockovčiaková, R. Soták, and R. Škrekovski. Star edge coloring of some classes of graphs. J. Graph Theory, 81:73–82, 2016. 13 [2] C. J. Casselgren, J. B. Granholm, and A. Raspaud. On star edge colorings of bipartite and subcubic graphs. 2019. ArXiv Preprint (https://arxiv.org/pdf/1912.02467.pdf). [3] Z. Dvořák, B. Mohar, and R. Šámal. Star chromatic index. J. Graph Theory, 72:313–326, 2013. [4] B. Grünbaum. Acyclic coloring of planar graphs. Israel J. Math., 14:390–412, 1973. [5] S. Kerdjoudj and A. Raspaud. List star edge coloring of sparse graphs. Discrete Appl. Math., 238:115–125, 2018. [6] H. Lei, Y. Shi, and Z.-X. Song. Star chromatic index of subcubic multigraphs. J. Graph Theory, 88:566–576, 2018. [7] H. Lei, Y. Shi, Z.-X. Song, and T. Wang. Star 5-edge-colorings of subcubic multigraphs. Discrete Math., 341:950–956, 2018. [8] X.-S. Liu and K. Deng. An upper bound on the star chromatic index of graphs with   7. J. Lanzhou Univ. (Nat. Sci.), 44:94–95, 2008. [9] B. Lužar, M. Mockovčiaková, and R. Soták. Note on list star edge-coloring of subcubic graphs. J. Graph Theory, 90:304–310, 2019. [10] M. Mockovčiaková. Distance Constrained Edge Colorings of Graphs. PhD thesis, P. J. Šafárik University, Faculty of Science, Košice, 2013. [11] B. Omoomi and M. V. Dastjerdi. Star Edge Coloring of the Cartesian Product of Graphs. 2018. Preprint (https://arxiv.org/abs/1802.01300). [12] J. J. Sylvester. On Subvariants, i.e. Semi-Invariants to Binary Quantics of an Unlimited Order. Amer. J. Math., 5:79–136, 1882. [13] Y. Wang, W. Wang, and Y. Wang. Edge-partition and star chromatic index. Appl. Math. Comput., 333:480–489, 2018. [14] Y. Wang, Y. Wang, and W. Wang. Star edge-coloring of graphs with maximum degree four. Appl. Math. Comput., 340:268–275, 2019. 14 A Cartesian products of two cycles Figure 1: Cartesian products of C with cycles 6 0 0 6 5 3 4 0 1 2 0 0 3 4 0 3 4 5 1 2 1 2 4 6 4 0 2 0 1 2 0 4 5 3 3 5 3 5 1 2 0 1 4 2 6 1 4 6 0 0 6 5 3 4 (a) A star 6-edge-coloring of C  C (b) A star 7-edge-coloring of C  C 3 3 3 4 4 3 6 4 6 0 1 2 0 4 6 5 0 5 1 2 3 0 4 3 5 4 6 5 2 4 3 0 5 1 2 1 2 0 1 0 1 0 4 1 6 3 0 4 5 4 3 5 4 2 3 5 6 2 4 5 1 2 0 1 2 3 0 6 5 0 1 0 2 5 3 5 4 3 5 1 4 3 4 3 6 4 6 0 1 2 0 4 6 5 (d) A star 7-edge-coloring of C  C including a star 3 7 (c) A star 7-edge-coloring of C  C 3 5 6-edge-coloring of C  C (darker vertices) 3 3 0 1 2 0 1 0 1 5 4 3 5 4 6 5 4 2 4 1 2 0 1 0 3 5 0 5 4 3 5 4 2 1 3 5 2 0 1 2 3 6 0 6 3 5 4 3 5 4 2 4 3 0 0 0 5 1 2 1 1 (e) A star 7-edge-coloring of C  C including a star 6-edge-coloring of C  C (darker vertices) 3 8 3 3 15 Figure 2: Cartesian products of C with cycles 6 0 0 1 1 0 2 4 1 4 0 5 4 3 2 0 2 4 1 0 3 5 2 3 3 6 1 4 3 5 3 5 2 0 0 1 2 0 5 0 5 2 1 4 0 3 5 2 0 3 4 6 4 0 2 1 4 1 5 4 5 0 2 5 2 1 4 0 5 3 2 0 6 1 6 0 5 3 5 3 2 0 3 1 2 5 4 3 3 2 4 1 0 5 3 0 0 0 1 1 0 2 4 1 4 (a) A star 7-edge-coloring of C  C (b) A star 6-edge-coloring of C  C 4 5 4 6 5 4 2 5 3 3 3 0 1 0 2 0 1 0 4 4 3 5 6 4 6 3 2 5 1 3 0 5 3 0 0 4 4 2 3 2 5 1 3 2 5 1 4 5 3 5 0 0 3 6 0 4 2 4 1 6 5 1 4 5 3 3 5 4 2 3 (c) A star 7-edge-coloring of C  C including a star edge-coloring of C  C (darker vertices) 4 7 4 4 5 3 5 3 4 0 4 0 3 0 1 0 2 0 1 3 2 6 0 4 4 3 5 6 2 1 4 5 3 2 5 1 0 1 5 0 6 3 0 0 4 4 3 4 2 1 2 5 1 3 2 5 1 0 3 4 5 3 5 0 0 6 5 6 2 0 4 2 4 1 5 2 3 5 1 4 5 3 5 3 4 0 4 0 3 (d) A star 7-edge-coloring of C  C including a star edge-coloring of C  C (darker vertices) 4 9 4 4 16 Figure 3: A star 6-edge-coloring of C  C combined of star edge-colorings of 4 10 C  C (lighter vertices) and C  C (darker vertices) 4 4 4 6 1 4 1 4 1 0 2 4 1 4 2 0 3 5 2 4 1 0 3 5 2 3 5 2 0 3 5 3 5 2 0 5 2 0 3 5 2 1 4 0 3 5 4 1 4 1 4 0 2 1 4 1 2 0 5 3 2 1 4 0 5 3 2 5 3 2 0 5 3 5 3 2 0 3 2 0 5 3 2 4 1 0 5 3 1 4 1 4 1 0 2 4 1 4 Figure 4: Cartesian products of C with cycles 6 5 6 4 6 5 4 0 1 3 4 1 3 0 1 2 0 3 4 6 3 4 5 1 2 4 1 4 3 6 5 0 1 5 0 6 3 0 2 6 2 4 1 2 3 2 0 3 1 2 6 3 5 0 0 5 0 1 6 1 2 4 0 5 1 3 1 4 2 1 6 4 5 6 3 1 6 6 5 6 3 4 0 3 2 0 6 3 1 3 1 4 1 0 3 4 5 1 4 1 2 5 4 2 2 5 2 5 2 0 3 3 0 6 3 4 0 3 0 1 4 6 1 2 6 2 5 0 6 6 5 6 4 6 5 4 0 1 3 4 1 (b) A star 7-edge-coloring of C  C , including a star 5 7 (a) A star 7-edge-coloring of C  C 5 5 edge-coloring of C  C (darker vertices) 5 3 2 5 3 1 5 6 3 1 3 0 5 0 1 2 0 3 1 2 0 5 1 4 0 6 4 5 2 4 5 5 2 6 6 3 1 0 3 1 3 6 4 1 3 2 0 1 3 2 4 5 0 3 3 4 5 5 4 2 5 1 2 3 1 0 2 0 4 6 2 4 3 0 6 2 5 6 3 1 3 0 5 1 6 4 1 4 2 5 0 4 6 5 6 0 2 5 6 0 0 6 2 2 2 4 3 1 4 0 3 1 4 0 5 1 4 2 3 1 5 6 3 1 3 0 5 (c) A star 7-edge-coloring of C  C 5 11 17 Figure 5: A star 7-edge-coloring of C  C including a star edge-coloring of 7 7 C C (darker vertices in horizontal direction) and a star edge-coloring of C C 3 7 7 3 (darker vertices in vertical direction) 0 1 0 0 4 3 0 3 4 2 3 2 1 2 3 1 6 6 6 5 5 6 3 0 5 1 4 0 2 3 2 3 4 2 3 4 4 5 6 2 5 1 6 1 5 1 4 0 3 0 3 0 6 3 5 6 4 5 2 6 0 1 2 0 2 1 4 3 5 4 3 1 6 3 2 2 6 5 0 2 6 1 6 0 1 4 5 3 1 4 3 4 6 3 0 2 5 6 2 5 2 6 4 5 0 1 0 0 4 3 0 18 Figure 6: A star 7-edge-coloring of C  C including a star edge-coloring of 11 11 C  C (all darker vertices), which furthermore includes a star edge-coloring of 7 11 C  C (the darkest vertices above) 3 11 0 1 2 5 2 5 6 2 4 2 4 2 6 3 4 0 1 4 0 5 6 5 2 1 0 5 6 3 6 3 1 3 4 3 5 4 2 0 4 5 5 0 4 6 2 0 2 3 1 3 1 2 1 0 0 1 1 3 5 4 0 6 4 0 3 5 3 6 3 0 1 2 2 0 5 6 2 1 2 2 1 6 3 5 1 3 0 1 4 0 5 1 3 0 4 3 2 4 2 0 2 3 0 4 2 5 0 6 3 0 6 5 1 6 4 1 3 1 4 5 1 5 3 3 4 3 5 6 5 2 3 6 2 1 4 2 0 5 2 4 0 1 1 3 3 6 0 1 1 3 5 3 6 0 4 3 0 5 1 3 6 0 0 2 5 2 5 6 2 4 2 2 1 6 1 3 0 1 3 0 1 3 0 1 3 2 4 1 3 4 2 3 6 5 4 1 5 0 6 2 6 1 5 4 0 6 1 2 3 5 4 5 0 3 2 1 4 5 0 6 1 2 1 3 6 4 0 5 2 0 4 4 6 0 0 2 1 6 3 1 3 1 5 0 3 6 4 0 5 1 0 6 1 0 1 2 5 2 5 6 2 4 2 4 19 B Cartesian products of paths and cycles Figure 7: Cartesian products of cycles and P 2 0 4 1 0 0 5 3 2 3 5 4 4 1 0 4 2 1 0 5 2 3 3 2 3 1 5 1 0 5 4 3 0 5 1 2 2 1 5 4 3 0 5 4 1 5 1 3 2 3 3 2 5 4 1 2 0 4 1 0 0 (a) A star 6-edge-coloring of C  P 5 6 3 1 5 3 1 5 0 2 4 0 2 5 3 1 5 3 1 2 4 0 2 4 1 5 3 1 5 3 4 0 2 4 0 3 1 5 3 1 5 3 4 0 5 3 5 0 2 4 0 2 5 3 2 1 0 4 3 1 5 3 3 1 1 4 0 2 4 2 4 0 1 4 0 2 5 3 0 0 5 3 2 5 1 4 3 1 4 5 1 2 1 0 4 2 2 5 3 0 2 3 4 5 1 0 3 0 4 0 2 1 3 2 0 4 5 2 3 5 1 4 0 5 1 2 3 1 4 2 1 2 3 3 2 0 5 4 0 2 3 4 0 5 1 3 3 1 5 3 0 4 0 5 1 2 0 2 4 0 2 4 5 3 4 0 3 1 5 3 1 5 3 1 2 1 2 5 4 4 0 2 4 0 0 5 3 4 1 3 1 5 3 1 5 3 4 0 5 3 5 (c) A star 6-edge-coloring of C  P including a 10 6 (b) A star 6-edge-coloring of C  P star edge-coloring of C  P (darker vertices) 7 6 3 6 20 Figure 8: Cartesian products of cycles and P 2 3 0 2 1 0 2 1 2 3 4 2 3 4 2 3 1 5 4 3 5 4 3 0 1 5 0 1 5 0 0 2 1 0 2 1 0 2 4 2 3 4 2 3 4 2 5 4 3 5 4 3 1 1 5 0 1 5 0 1 1 0 2 1 0 2 5 0 3 4 2 3 4 2 3 4 4 3 5 4 1 0 4 5 0 1 5 0 1 5 2 1 0 2 3 3 2 3 1 3 4 2 3 4 2 3 3 5 1 0 5 4 1 0 1 5 0 1 5 4 0 2 4 4 2 0 3 2 2 2 3 4 2 3 0 0 5 0 3 5 1 2 5 4 5 0 1 4 2 1 1 1 2 0 4 4 0 4 3 1 2 3 5 5 3 2 3 4 5 2 3 5 1 5 0 1 2 0 1 0 0 0 1 1 0 2 4 0 2 3 4 4 3 2 5 4 5 2 3 4 5 0 3 0 2 5 0 1 3 0 2 4 0 2 1 1 2 2 1 1 3 2 4 4 2 3 1 3 5 0 3 4 5 5 4 0 3 5 0 1 0 2 4 4 2 0 1 1 2 2 1 1 2 3 4 2 3 0 1 3 5 2 3 0 3 5 4 0 2 5 4 5 2 0 4 4 0 2 3 4 2 3 1 1 3 3 1 4 3 2 1 3 5 1 5 0 2 5 4 0 5 0 5 5 0 2 1 0 2 3 4 4 3 2 1 1 4 2 1 4 3 5 4 5 2 1 5 0 3 5 2 3 0 2 1 0 2 1 0 0 3 2 4 4 2 3 4 1 5 3 1 5 0 (a) A star 6-edge-coloring of C  P 11 8 3 2 0 0 2 3 1 2 1 3 4 1 5 0 4 5 5 2 3 4 2 3 1 4 0 1 5 0 1 5 2 3 4 2 3 4 2 3 (b) A star 6-edge-coloring of C  P including a star 14 8 edge-coloring of C  P (darker vertices) 3 8 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Mathematics arXiv (Cornell University)

Star Edge-Coloring of Square Grids

Download PDF
21 pages

Loading next page...
 
/lp/arxiv-cornell-university/star-edge-coloring-of-square-grids-GROLSinWt8

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

ISSN
0096-3003
eISSN
ARCH-3343
DOI
10.1016/j.amc.2020.125741
Publisher site
See Article on Publisher Site

Abstract

A star edge-coloring of a graph G is a proper edge-coloring without bichromatic paths or cycles of length four. The smallest integer k such that G admits a star edge-coloring with k colors is the star chromatic index of G. In the seminal paper on the topic, Dvořák, Mohar, and Šámal asked if the star chromatic index of complete graphs is linear in the number of vertices and gave an almost linear upper bound. Their question remains open, and consequently, to better understand the behavior of the star chromatic index, this parameter has been studied for a number of other classes of graphs. In this paper, we consider star edge-colorings of square grids; namely, the Cartesian products of paths and cycles and the Cartesian products of two cycles. We improve previously established bounds and, as a main contribution, we prove that the star chromatic index of graphs in both classes is either 6 or 7 except for prisms. Additionally, we give a number of exact values for many considered graphs. Keywords: star edge-coloring, star chromatic index, square grid, Cartesian product. 1 Introduction A proper edge-coloring of a graph G is called a star edge-coloring if there is neither bichromatic path nor bichromatic cycle of length four. The minimum number of colors for which G admits a star edge-coloring is called the star chromatic index and we denote it by  (G). st The star edge-coloring was defined in 2008 by Liu and Deng [8], and was motivated by the vertex version introduced by Grünbaum [4]. Despite a number of papers have already been published about this coloring, we have a very limited knowledge about it. In particular, the exact value of the star chromatic index of complete graphs is still not known, although some relatively strong lower and upper bounds have been determined by Dvořák et al. in their seminal paper [3]. Theorem 1 (Dvořák, Mohar, Šámal, 2013). The star chromatic index of the complete graph K satisfies p p 2 2(1+o(1)) log n 2n(1 + o(1))   (K )  n : st 1=4 (log n) 0 1+ In particular, for every  > 0 there exists a constant C such that  (K )  C n for every st n  1. Faculty of Applied Sciences, University of West Bohemia, Pilsen, Czech Republic. E-Mails: holubpre@kma.zcu.cz, mmockov@ntis.zcu.cz Faculty of Information Studies, Novo mesto, Slovenia. E-Mail: borut.luzar@gmail.com Faculty of Science, Pavol Jozef Šafárik University, Košice, Slovakia. E-Mails: erika.mihalikova@student.upjs.sk, roman.sotak@upjs.sk arXiv:2005.02864v2 [math.CO] 19 Aug 2020 They proved the upper bound using a nontrivial result about sets without arithmetic progressions, and up till now, it is still the best known. For the lower bound, they used an elegant double counting approach. The authors of [1] observed a small improvement in their proof and obtained the bound  (K )  3n(n 1)=(n + 4) (see [10] for a proof), which gives the exact values for st the chromatic index of K , for n 2 f1; 2; 3; 4; 8g. However, despite all efforts, the asymptotic behavior of the star chromatic index of complete graphs is not known, and in [3] the following question has been asked. Question 1 (Dvořák, Mohar, Šámal, 2013). What is the true order of magnitude of  (K )? st Is  (K ) = O(n)? st Another class of graphs with highly regular structure are complete bipartite graphs. They are important for better understanding of the coloring already on their own, and also, as Dvořák et al. [3] observed, the bounds for their star chromatic index provide bounds for the index of complete graphs. dlog ne 0 0 i2 0 (K ) n   (K )  2  (K i i ) : n;n n dn=2 e;dn=2 e st st st i=1 Recently, Casselgren et al. [2] considered complete bipartite graphs and proved the tight upper bound for K ; r  5, derived a lower and upper bound for K ; s  4, and, using computer, 3;r 4;s they also determined the star chromatic index for some complete bipartite graphs of small order. Star edge-coloring has been studied also for other classes of graphs, e.g., graphs with max- imum degree 3 [3, 6, 7] and 4 [14], subcubic Halin graphs [2], outerplanar graphs [1, 13], and planar graphs with various constraints [13]. Moreover, the list version of the star edge-coloring has also been investigated (see, e.g., [5, 9]). Finally, there is also a complexity result on the topic; namely, it is NP-complete to decide whether 3 colors suffice for a star edge-coloring of a subcubic multigraph [6]. Since most of the obtained upper bounds for the star chromatic index are not tight and many questions remain open, we focus our attention to graphs with a relatively simple structure, i.e. to the Cartesian products of graphs. The star edge-coloring of the Cartesian products of graphs has already been considered by Omoomi and Dastjerdi [11]. They established an upper bound for the star chromatic index of the Cartesian product of two arbitrary graphs, proved its exact values for the Cartesian product of two paths (Theorem 4), and they started investigation on the Cartesian products of a path and a cycle, and the Cartesian product of two cycles (i.e., square grids). They further proved upper bounds for d-dimensional grids and d-dimensional hypercubes. Motivated by the results presented in [11], in this paper, we consider star edge-coloring of square grids; in particular, the Cartesian products of two cycles and the Cartesian products of paths and cycles. Apart from the usual combinatorial methods, due to the complexity of the problems considered in this paper, we have used computer to obtain star edge-colorings of small graphs and to establish some of the lower bounds. Standard (formal) mathematical proofs would require enourmous amount of case analysis, while their contribution to the theory would be minimal. We establish exact bounds for the star chromatic index of many graphs from the two considered classes, and show that the upper bound for the chromatic index of both Cartesian products is 7. The paper is structured as follows. We give our notation and prove some auxiliary results in Section 2. Section 3 contains the algorithm used in our computations and describes the 2 preprocessing procedures used in them. In Section 4, we present the main results of this paper, and we list some open problems in Section 5. 2 Preliminaries In this section, we present some additional terminology used in the paper and give auxiliary results. We abbreviate a ‘star edge-coloring with k colors’ to a ‘star k-edge-coloring’, and, if it is clear from the context, sometimes we just write ‘coloring’ instead of ‘star edge-coloring’. The Cartesian product of graphs G and H, denoted by G  H, is the graph with the vertex 0 0 set V (G) V (H ) and edges between the vertices (u; v) and (u ; v ) if: 0 0 uu 2 E(G) and v = v (a G-edge), or 0 0 u = u and vv 2 E(H ) (an H-edge). We call the graphs G and H the factor graphs. The G-fiber with respect to v 2 V (H ), denoted by G , is the copy of G in G  H induced by the vertices having v as the second component. Analogously, the H-fiber with respect to u 2 V (G), denoted by H , is the copy of H in G  H induced by the vertices having u as the first component. Since the Cartesian product of two paths is a subgraph of the Cartesian product of a path and a cycle, and the Cartesian product of a path and a cycle is a subgraph of the Cartesian products two cycles, we have the following sequence of inequalities. Observation 1. For every pair of positive integers m and n, where m  3 and n  3, we have 0 0 0 (P  P )   (C  P )   (C  C ): m n m n m n st st st Having a star edge-coloring of the Cartesian product of an n-cycle and a graph H, we can extend it to a coloring of the Cartesian product of a cycle of length k  n and H. Lemma 1. For every integers k and m, where k  2 and m  3, and for every graph H, we have 0 0 (C  H )   (C  H ) : km m st st Proof. Let C = u : : : u u , C = v : : : v v , and V (H ) = fw ; : : : ; w g. Moreover, let m 1 m 1 km 1 km 1 1 n p : f1; : : : ; kmg ! f1; : : : ; mg be an assignment given by p(t) = s if and only if (ts) is divisible by m. Let  be a star edge-coloring of C  H. Consider an edge e = (v ; w )(v ; w ) of C  H. m i a j b km By the definition of the Cartesian product, we have i = j or a = b. Note that if a = b, then ji jj = 1, and thus we may assume j = i + 1. We define a proper edge-coloring  of C  H km as follows. If i = j, then set  ((v ; w )(v ; w )) = ((u ; w )(u ; w )). In the case a = b, we i a i b a b p(i) p(i) set  ((v ; w )(v ; w )) = ((u ; w )(u ; w )). i a j a p(i) a p(j) a Now we show that  is also a star edge-coloring. For an integer s, where 1  s  k  m, let G be the graph induced by the vertices f(v ; w)g, where ` 2 fs + 1; : : : ; s + mg (the values s ` s + 1; : : : ; s + m are taken modulo k  m) and all w 2 V (H ), i.e., G is the graph induced on m consecutive H-fibers. Observe that the coloring  on G corresponds to a coloring  of the subgraph of C H without the edges (u ; w)(u ; w), for all w 2 V (H ). Therefore, p(s+m) p(s+1) every 4-path and every 4-cycle in C  H, contained in some G , is not bichromatic. km s Finally, if a 4-path or a 4-cycle is not contained in any G , then it contains at least m edges of type (v ; w )(v ; w ) (i.e., only when m 2 f3; 4g). However, in the case of m = 3, i a j a 3 three consecutive edges on every C -fiber receive three distinct colors, and hence no 4-path with three consecutive edges on a C -fiber is bichromatic. If a 4-path has two consecutive edges on a C -fiber, an edge in an H-fiber, and the fourth edge in another C -fiber, then its coloring m m corresponds to a coloring of some 4-path in C  H, which is not bichromatic. In the case of m = 4, we only have 4-cycles, whose colorings correspond to a coloring of a 4-cycle by  in some C -fiber, and hence they are not bichromatic. We continue by showing how star edge-colorings of two Cartesian products, each having at least one cycle as a factor, can be combined. Let m and n be a pair of integers, where 3  m < n, and let v ; : : : ; v be consecutive vertices of the cycle C . We say that a star edge-coloring  of 1 n n C  H includes a star edge-coloring of C  H if the coloring  of the subgraph of C  H n m n induced by the vertices of m consecutive H-fibers H ; : : : ; H , together with the additional v v 1 m edges e = (v ; w)(v ; w), for all w 2 V (H ), where we set  (e ) = ((v ; w)(v ; w)), is a star w 1 m w 1 n edge-coloring. Symmetrically, we can say that a star k-edge-coloring of H C includes a star edge-coloring of H  C . Note that the star edge-coloring of C  H, constructed in the proof of Lemma 1, m km includes a star edge-coloring of C  H. Lemma 2. If for a pair of positive integers m and n, where m < n, a star k-edge-coloring of C  H includes a star edge-coloring of C  H, then, for every pair of non-negative integers p n m and q, we have (C  H )  k : pm+qn st Proof. Let  be a star k-edge-coloring of C H which includes a star edge-coloring  of C H. n m Let C = v : : : v v , C = v : : : v v , and C = u : : : u u . Furthermore, we define n 1 n 1 m 1 m 1 pm+qn 1 pm+qn 1 an assignment r : f1; : : : ; pm +qng ! f1; : : : ; ng such that, if t  pm, then r(t)  m and tr(t) is divisible by m, and, if t > pm, then t pm r(t) is divisible by n. Now, similarly as in the proof of Lemma 1, we define an edge-coloring  of C  H. We pm+qn combine p copies of  followed by q copies of . More precisely, for w ; w 2 H, a b ((v ; w )(v ; w )); for i = j  pm ; > a r(i) r(i) b ((v ; w )(v ; w )); for i = j > pm ; r(i) a r(i) b ((u ; w )(u ; w )) = i a j b ((v ; w )(v ; w )); for i = j 1  pm ; > a a r(i) r(j) ((v ; w )(v ; w )); for i = j 1 > pm : a a r(i) r(j) Note that, by the definition of  , we have  ((v ; w)(v ; w)) = ((v ; w)(v ; w)). m 1 n 1 It remains to show that  is a star edge-coloring. For an integer s, where 1  s  pm + qn, let G be the graph induced by the vertices of m + 1 consecutive H-fibers H ; : : : ; H s u u s+1 s+m+1 (the indices s + i are taken modulo pm + qn). Since the coloring of each G is a part of two consecutive copies of  or a part of two consecutive copies of , no 4-path and no 4-cycle in G is bichromatic (using Lemma 1 for k = 2). Finally, if a 4-path is not contained in any G , then it contains 4 edges of type (v ; w )(v ; w ) s i a j a and m = 3. Moreover, such a 4-path traverses the H -fibers, for i 2 fpm; : : : ; pm + 4g or i 2 fpm + qn; 1; : : : ; 4g. In both cases, colors of three consecutive edges of the 4-path correspond to colors of a C -fiber of  and therefore they are distinct. This completes the proof. We will use Lemma 2 to prove results for arbitrary lengths of cycles. To do that, we will use the following result on Frobenious numbers [12]. 4 Theorem 2 (Sylvester, 1882). Let positive integers n and m be relatively prime. Then for every integer k  (n 1)(m 1) there exist non-negative integers and such that k =  n +  m : We also recall the result of Dvořák, Mohar, and Šámal [3] about star edge-coloring of subcubic graphs, which we will use when considering prisms. Theorem 3 (Dvořák, Mohar, Šámal, 2013). (a) If G is a subcubic graph, then  (G)  7. st (b) If G is a simple cubic graph, then  (G)  4, and the equality holds if and only if G covers st the graph of the 3-dimensional hypercube. Here, a graph G is said to cover a graph H if there is a graph homomorphism from G to H that is locally bijective. In other words, there is a mapping f : V (G) ! V (H ) such that whenever uv is an edge of G, the image f (u)f (v) is an edge of H, and, for each vertex v 2 V (G), f is a bijection between the neighbors of v and the neighbors of f (v). At this point, we remark the following, somehow hidden, corollary of the above result. Hexag- onal grids are subcubic graphs and they cover the graph of the 3-dimensional hypercube. Thus: Corollary 1. For an infinite hexagonal grid G, we have (G) = 4 : st 3 Computer Computations and Algorithm For our computations, we used a simple backtracking algorithm, which, together with some preprocessing, enabled us to compute exact lower bounds for some important cases on one hand, and on the other hand, provided star edge-colorings with required properties for some graphs. The main coloring algorithm takes three input parameters: the graph to be colored, the number of colors, and a possible precoloring of some edges, in order to avoid testing some isomorphic partial colorings; e.g., one may fix the colors on the edges incident to a vertex of maximum degree. Note also that before calling the function StarColor, we first verify that the precoloring of the edges is a star edge-coloring. Another important part of our algorithm is determining the order of edges (the function GetEdgeOrdering), in which it tries to color them. We order the edges (ignoring the precolored edges) by the number of precolored neighbors (incident edges) and the number of neighbors appearing earlier in the ordering in a descending order. The function Conflict checks if assigning a color to the current edge introduces a conflict, namely, it checks if two adjacent edges receive the same color, and if a bichromatic 4-path or 4-cycle appears. In some cases, we manually controlled the different cases of precolored edges. If we established some additional property of a required coloring, e.g., that no 4-path can be colored with just three colors, we included that in the procedure. Finally, we also adopted the algorithm to output all possible colorings of a given graph, and in the case of symmetric graphs, e.g., cycles, we eliminated isomorphic colorings. The remaining colorings were used to test if they can be extended to graphs on more vertices. 5 Algorithm 1 Star edge-coloring algorithm 1: procedure StarColor(G; k;P) . Graph G, number of colors k, precolored edges P 2: edgeOrder GetEdgeOrdering(G, P) 3: edgeColors InitEdgeColors(P,edgeOrder) 4: triedColors InitTriedColors(edgeOrder) . Dictionary of empty lists for all edges 5: for i in 1::edgeOrder.Count do . Try to color edges according to the ordering 6: e edgeOrder[i] 7: isColored false 8: for color c in f1::kg n triedColors[e] do 9: edgeColors[e] c 10: if Conflict(c, G, edgeColors) then . Check if a conflict occurs 11: edgeColors[e] ; 12: else 13: add c to triedColors[e] 14: isColored true 15: goto 18 16: end if 17: end for 18: if !isColored and i > 1 then . If no color is found, continue if not at first edge 19: edgeColors[edgeOrder[i 1]] ; 20: i = i 2 . Step up, 2 handles automatic loop increment 21: else if !isColored and i == 1 then 22: return “No coloring found” 23: else if isColored and i == edgeOrder.Count then 24: return edgeColors . A star edge-coloring is found 25: end if 26: end for 27: end procedure 4 Cartesian products of paths and cycles 4.1 Cartesian products of paths In a recent paper, Omoomi and Dastjerdi [11] established tight bounds for the star chromatic index of two paths (see Table 1). Theorem 4 (Omoomi and Dastjerdi, 2019). For the graph P P , where m and n are integers m n with 2  m  n, we have 3; if m = n = 2; 4; if m = 2 and n  3; (P  P ) = m n st 5; if m = 3 and 3  n  4; 6; otherwise. As a corollary, we establish the lower bound of 6 colors for the Cartesian products, where one factor is a cycle and the other is a path of length at least 2. 6 + mnn 2 3 4 5 2 3 4 4 4 3 4 5 5 6 4 4 5 6 6 5 4 6 6 6 Table 1: The star chromatic index of the Cartesian products of two paths (P  P ) m n st Corollary 2. For every pair of integers m and n, where m  3 and n  3, we have (C  P )  6: m n st Proof. We first note that the graph C P is one of the two known examples of simple bridgeless 3 2 cubic graphs that have star chromatic index equal to 6 [9]. Then, for m = 3, by Observation 1, 0 0 we have  (C  P )   (C  P ) = 6. 3 n 3 2 st st If m = 4, we proceed by a contradiction. Suppose that  (C  P )  5. Then, by Lemma 1, 4 n st 0 0 (C  P )  5, for any integer `, and hence also  (P  P )  5, a contradiction. Finally, n 5 n 4` st st 0 0 if m  5, then we have  (C  P )   (P  P ) = 6 by Theorem 4 and Observation 1. m n 5 3 st st 4.2 Cartesian products of cycles Having the Cartesian products of paths resolved, the logical direction of research is consideration of cylinders and toroidal grids, i.e., the Cartesian products of cycles and paths, and the Cartesian products of two cycles. We begin by giving some results about the latter. Corollary 2 implies that the Cartesian product of any two cycles will need at least 6 colors for a star edge-coloring. On the other hand, as we will show in this subsection, the star chromatic index of the Cartesian product of two cycles is at most 7. We first investigate the Cartesian products of C with another cycle. Theorem 5. For every integer n, where n  3, we have 6; if n = 3k, (C  C ) = 3 n st 7; otherwise Proof. By Corollary 2,  (C  C )  6. Now suppose that n = 3k for some integer k  1. If 3 n st k = 1, then there is a star 6-edge-coloring of C  C (one is depicted in Figure 1(a)). Next, by 3 3 Lemma 1, we have  (C  C ) = 6. 3 n st Using Algorithm 1, we infer that the Cartesian product C  P has only one star 6-edge- 3 3 coloring up to a permutation of colors. Namely, three colors, say 0, 1, and 2, appear on C -fibers, and the colors 4, 5, and 6 on the P -fibers. Since C  P is a subgraph of every graph C  C , 3 3 3 3 n it follows that such Cartesian products admit a star 6-edge-coloring only when n is divisible by Therefore, if n 6= 3k for every integer k, then  (C  C )  7. In Figures 1(d) and 1(e), 3 n st a star 7-edge-coloring of C  C and C  C , respectively, is depicted. Observe that, in both 3 7 3 8 colorings, a star 6-edge-coloring of C  C is included. Hence, by Lemma 2 and Theorem 2, we 3 3 have  (C  C ) = 7 for every n, n  7, not divisible by 3. The remaining two cases, namely 3 n st n = 4 and n = 5, are depicted in Figures 1(b) and 1(c), respectively. 7 Similarly as in the proof of Theorem 5, we can use Lemma 1 (twice) to extend the star edge-coloring of C  C to products of cycles of lengths divisible by 3. 3 3 Corollary 3. For every pair of positive integers k and `, we have (C  C ) = 6 : 3k 3` st We proceed with a result about the Cartesian products of C with another cycle. Theorem 6. For every pair of positive integers k and `, where k  1 and `  2, we have (C  C ) = 6 : 4k 2` st Proof. We use the star 6-edge-coloring  of C  C depicted in Figure 3. Note that 10 4 10 10 includes a star 6-edge-coloring C  C , and a star 6-edge-coloring C  C . Therefore, by 4 4 4 6 Lemma 2 and Theorem 2, we have  (C  C ) = 6 for every integer `  4. Finally, we use 4 2` st Lemma 1 to infer  (C  C ) = 6 for every integer k. 4k 2` st Proposition 1. For n 2 f5; 7; 9; 11g, we have (C  C ) = 7 : 4 n st Proof. Using Algorithm 1, we established that  (C  C ) > 6 for every n 2 f5; 7; 9g. The 4 n st bounds  (C  C ) = 7 follow from the star 7-edge-colorings depicted in Figures 2(a), 2(c), 4 n st and 2(d). In the case of n = 11, we split the computation in two steps. First, using Algorithm 1, we determined that if the edges of some C -fiber are colored in such a way that a same color appears twice on some 4-path, then the coloring cannot be extended to a star 6-edge-coloring of C  C . In the second step, the algorithm checked only the colorings in which every 4-path 4 11 in each C -fiber had four colors on its edges. It turned out that such a coloring does not exist. Therefore,  (C  C ) = 7 by the star edge-coloring depicted in Figure 2(c) and Lemma 1. 4 11 st Theorem 7. For any odd integer n, where n  13, we have (C  C )  7 : 4 n st Proof. In Figure 2(c), we present a star 7-edge-coloring of C C with a star 7-edge-coloring of 4 7 C  C included. Thus, by Lemma 2 and Theorem 2, we infer that  (C  C )  7 for every 4 4 4 n st odd n, where n > 18. Colorings for n 2 f13; 15; 17g can be obtained by using Lemma 2 and the colorings depicted in Figures 2(a) (for n = 15) and 2(d) (for n 2 f13; 17g). Theorem 8. For every integer n, where n  3, we have (C  C ) = 7 : 5 n st Proof. The lower bounds  (C  C ) > 6 for 3  n  6, were established using Algorithm 1. 5 n st For n  7, using Algorithm 1, we infer that  (C  P )  7. Therefore, by Observation 1, we 5 n st have  (C  C )  7. 5 n st Star 7-edge-colorings of C  C , for m 2 f3; 4; 5; 7; 11g, are depicted in Figures 1(c), 2(a), 5 m 4(a), 4(b), and 4(c), respectively. By Lemma 1, we also infer star 7-edge-colorings of C  C 5 m for m 2 f6; 8; 9; 10g. Finally, since in Figure 4(b), a star 7-edge-coloring of C  C is included, 5 3 by Lemma 2 and Theorem 2, we obtain  (C  C ) = 7 for every integer n  12. 5 n st 8 Theorem 9. For every integer n, where n  3, we have 6; if n  0 mod 3 or n  0 mod 4, (C  C ) = 6 n st 7; otherwise. Proof. By the star 6-edge-colorings depicted in Figures 1(a) and 2(b), and by Lemma 1, we have (C  C ) = 6, for every integer n divisible by 3 or 4. 6 n st Now we show that  (C  C ) > 6 if n is not divisible by 3 or 4. First, we consider the 6 n st graph C P , where P = v : : : v . We start with a precolored C -fiber at the vertex v (i.e., 6 31 31 1 31 6 16 the middle C -fiber) using each of the nine possible star 6-edge-colorings of C (up to symmetries 6 6 and permutations of colors). Using Algorithm 1, we tried to extend such a precoloring to the whole C  P . For five colorings of the C -fiber, namely for (0; 1; 0; 2; 0; 3), (0; 1; 0; 2; 1; 2), 6 31 6 (0; 1; 0; 2; 1; 3), (0; 1; 2; 0; 1; 3), and (0; 1; 2; 0; 3; 4), we obtain that such precolorings cannot be extended. For the remaining four precolorings, namely (0; 1; 0; 2; 3; 2), (0; 1; 0; 2; 3; 4), (0; 1; 2; 0; 1; 2), and (0; 1; 2; 3; 4; 5), we obtain 27 078 colorings of C  P in total. Some of them are either 4-, 6 31 or 6-periodical, i.e., the initial coloring repeats on every 4-th or 6-th fiber, except at the final three fibers on both sides, where the coloring restrictions are relaxed. The remaining 26 448 colorings correspond to the precoloring (0; 1; 2; 3; 4; 5), and moreover, all C -fibers are colored by shifts of this precoloring, and every pair of adjacent C -fibers is either 6 6 colored with the same sequence of colors, or the coloring of one is the coloring of the other shifted by 1. In a more detailed analysis of these colorings, we find that, if they are periodic, then the period must be a multiple of 6. Thus, for the graphs C  C , it follows that they are star 6-edge-colorable if n is divisible 6 n by 3 or 4. Otherwise they are not star 6-edge-colorable. Theorem 10. For every integer n, where n  3, we have (C  C ) = 7 : 7 n st Proof. The lower bounds  (C  C ) > 6, for n 2 f3; 4; 5; 6g, were established using Al- 7 n st gorithm 1. For n  7, using Algorithm 1, we infer that  (C  P )  7. Therefore, by 7 n st Observation 1, we have  (C  C )  7. 7 n st Star 7-edge-colorings of C  C , for n 2 f3; 4; 5; 7g, are depicted in Figures 1(d), 2(c), 4(b), 7 n and 5, respectively. By Lemma 1, from these colorings, we also infer star 7-edge-colorings of C  C for n 2 f6; 8; 9; 10g. Moreover, in the coloring depicted in Figure 6, a star 7-edge- 7 n coloring of C  C is included, and hence we also have  (C  C ) = 7. Finally, since in the 7 11 7 11 st coloring depicted in Figure 5, a star 7-edge-coloring of C  C is included, by Lemma 2 and 7 3 Theorem 2, we have  (C  C ) = 7 for every n  12. 7 n st Proposition 2. For C  C , we have 8 9 (C  C ) = 7 : 8 9 st Proof. We determined that  (C  C )  7 by exhaustive computer search. Namely, we gener- 8 9 st ated all 147 non-isomorphic star 6-edge-colorings of C and tried to extend each of them to the graph C  C . None of them could be extended, thus  (C  C )  7. The equality follows 8 9 8 9 st from Lemma 1 and the fact that  (C  C ) = 7. 4 9 st 9 Finally, we give a general result, showing that 7 is the upper bound for the star chromatic index of the Cartesian products of any two cycles. Theorem 11. For every pair of positive integers m and n, where 3  m  n, we have (C  C )  7 : m n st Proof. By Theorems 5-10, we have  (C  C )  7 for 3  m  7 and n  3. Furthermore, m n st by Lemma 1, we can use Theorem 7 to obtain a star 7-edge-coloring of C  C , Theorem 5 8 n to obtain a star 7-edge-coloring of C  C , and Theorem 8 to obtain a star 7-edge-coloring of 9 n C  C , for every n  3. The star 7-edge-coloring of C  C is depicted in Figure 6. 10 n 11 11 We complete the proof by showing that  (C  C )  7 if m; n  12. Note that the star m n st 7-edge-coloring of C C depicted in Figure 5, includes a 7-edge-coloring of C C and a 7-edge- 7 7 3 7 coloring of C C . Furthermore, the latter two colorings include a common star 7-edge-coloring 7 3 of C  C . This fact enables us to use Lemma 2 and Theorem 2 to obtain  (C  C )  7 3 3 m n st for m; n  12. The above results are summarized in Table 2. mnn 3 4 5 6 7 8 9 10 11 12 3 6 7 7 6 7 7 6 7 7 6 4 7 6 7 6 7 6 7 6 7 6 5 7 7 7 7 7 7 7 7 7 7 6 6 6 7 6 7 6 6 7 7 6 7 7 7 7 7 7 7 7 7 7 7 8 7 6 7 6 7 6 7 6 7 6 9 6 7 7 6 7 7 6 7 7 6 10 7 6 7 7 7 6 7 7 7 6 11 7 7 7 7 7 7 7 7 7 7 12 6 6 7 6 7 6 6 6 7 6 Table 2: The star chromatic index of the Cartesian products of cycles (C  C ). In red, we denote the cases, where the exact bounds are not estab- m n st lished yet. The value 7 means that the exact value of the star chromatic index is either 6 or 7. 4.3 Cartesian products of cycles and paths In the last part of this section, we give results about the Cartesian products of paths and cycles. We begin with proving the cases for specific lengths of cycles. Theorem 12. For every pair of integers k and n, where k  2 and n  3, we have (C  P ) = 6 : 2k n st Proof. By Corollary 2, we have  (C  P )  6. On the other hand, 6 colors also suffice by 2k n st Theorem 6, since C  P is a subgraph of C  C for every `  n=4. 2k 2k 4` 10 Theorem 13. For every pair of integers k and n, where n  3, we have (C  P ) = 6 : 3k n st Proof. By Corollary 2, we have  (C  P )  6. On the other hand, 6 colors also suffice by 3k n st Corollary 3, since C  P is a subgraph of C  C for every `  n=3. 3k n 3k 3` Theorem 14. For every integer n, where n  3, we have 6; if n 2 f3; 4; 5; 6g, (C  P ) = 5 n st 7; if n  7. Proof. Let G = C P for some integer n  3. Suppose first that n 2 f3; 4; 5; 6g. By Theorem 4, 5 n 0 0 we have  (P  P ) = 6, and thus, since P P is a subgraph of G, it follows that  (G)  6. 3 5 5 3 st st On the other hand, in Figure 7(a), we give a star 6-edge-coloring of C  P , hence establishing 5 6 (C  P ) = 6 for every n 2 f3; 4; 5; 6g. Now, suppose that n  7. Using Algorithm 1, we 5 n st 0 0 infer that  (C  P )  7. The upper bound  (C  P )  7 follows from the fact that 5 n 5 n st st (C  C ) = 7 for every positive integer k (see Theorem 8 and Figure 4(a)). 5 5k st Theorem 15. For every integer n, where n  3, we have 6; if n 2 f3; 4; 5; 6g, (C  P ) = 7 n st 7; if n  7. Proof. Let G = C  P for some integer n  3. Suppose first that n 2 f3; 4; 5; 6g. Since 7 n 0 0 (P  P ) = 6, we again have  (G)  6. On the other hand, in Figure 7(b), we give a 3 5 st st star 6-edge-coloring of G, and hence  (G) = 6 for every n 2 f3; 4; 5; 6g. Suppose now that st 0 0 n  7. Using Algorithm 1, we infer that  (C  P ) = 7, and hence  (G)  7. The equality 7 7 st st is established by Figure 1(d), where a pattern for a star 7-edge-coloring of C C is presented. 7 3k Since G is a subgraph of C  C , for k large enough, the statement follows. 7 3k We now turn our attention to the Cartesian products of cycles and paths P , for m 2 f2; 3; 4g. Theorem 16. For every integer m, where m  3, we have >6; if m = 3, (C  P ) = 4; if m  0 mod 4, m 2 st 5; otherwise. Proof. For m = 3, recall that  (C  P ) = 6. If m  0 mod 4, then the graph C  P covers 3 2 m 2 st the graph Q , and hence its star chromatic index equals 4 by Theorem 3. Finally, if m 6 0 mod 4, by Theorem 3, we have  (C  P )  5. The equality follows from Theorem 7 in [11]. m 2 st Theorem 17. For every pair of integers m and n, where m  3 and n 2 f3; 4; 5; 6g, we have (C  P ) = 6 : m n st 11 Proof. First, recall that 6 colors are needed in all the cases by Corollary 2. Next, since C  P m 6 contains all the graphs C  P , for n 2 f3; 4; 5g, it suffices to show that there exists a star m n 6-edge-coloring of C  P . m 6 0 0 By Lemma 1, Theorems 5 and 6, we have  (C  P ) =  (C  P ) = 6, for any positive 3k 6 4k 6 st st integer k. Similarly, the star 6-edge-colorings depicted in Figures 7(a) and 7(b) together with 0 0 Lemma 1 imply  (C  P ) =  (C  P ) = 6. Furthermore, using the star 6-edge-coloring 5k 6 7k 6 st st of C  P depicted in Figure 7(c), which includes a star 6-edge-coloring of C  P , we obtain 10 6 3 6 (C  P ) = 6 for every m 2 f10; 13; 16g. Moreover, together with Lemma 2 and Theorem 2, m 6 st this coloring implies  (C  P ) = 6 for every m  18. For the remaining two cases, namely m 6 st m 2 f11; 17g, the corresponding colorings are depicted in Figures 8(a) and 8(b) (note that, in fact, we have even more, namely, we give a star 6-edge-coloring of C  P and a star 6-edge- 11 8 coloring of C  P which includes a star 6-edge-coloring of C  P ). 14 8 3 8 We collect known and our new results in Table 3. mnn 2 3 4 5 6 7 8 9 3 6 6 6 6 6 6 6 6 4 4 6 6 6 6 6 6 6 5 5 6 6 6 6 7 7 7 6 5 6 6 6 6 6 6 6 7 5 6 6 6 6 7 7 7 8 4 6 6 6 6 6 6 6 9 5 6 6 6 6 6 6 6 10 5 6 6 6 6 6 6 6 11 5 6 6 6 6 6 6 7 12 4 6 6 6 6 6 6 6 13 5 6 6 6 6 7 7 7 14 5 6 6 6 6 6 6 6 15 5 6 6 6 6 6 6 6 16 4 6 6 6 6 6 6 6 17 5 6 6 6 6 6 6 7 18 5 6 6 6 6 6 6 6 19 and m  0 mod 4 4 6 6 6 6 6 6 6 19 and m  r mod 12, r 2 f2; 3; 6; 9; 10g 5 6 6 6 6 6 6 6 19 and m  r mod 12, r 2 f1; 5; 7; 11g 5 6 6 6 6 7 7 7 Table 3: The star chromatic index of the Cartesian products of cycles and paths (C  P ). In red, we denote the cases, where the exact bounds are not estab- m n st lished yet. The value 7 means that the exact value of the star chromatic index is either 6 or 7. 5 Conclusion In this paper, we have established tight upper bounds for the Cartesian product of cycles and paths, and the Cartesian products of two cycles. We proved that 7 colors always suffice for 12 star edge-colorings of these graphs, which is, in a way, a surprising bound, especially because at least 6 colors are always needed as soon as one of the factors is not isomorphic to P . Although we improved existing bounds and proved a number of exact values, there are still some open questions. We are very confident that the following conjecture is true. Conjecture 1. There exist constants K and K such that for every pair of integers m and n, 1 2 where m  K and n  K , we have 1 2 (C  P ) = 6 : m n st Note that Conjecture 1 is equivalent to the following. Conjecture 2. There exists a constant K such that for every integer m, where m  K, there exists an integer n such that (C  C ) = 6 : m n st In fact, we believe (with a bit lower confidence) that the following stronger version of Con- jecture 1 can also be confirmed. Conjecture 3. There exists a constant L such that for every pair of integers m and n, where m; n  L, we have (C  C ) = 6 : m n st The above conjectures seem to be challenging, since we were not able to observe any straight- forward pattern in colorings of different Cartesian products, despite the help of computer in our constructions. There are also some (maybe) less complicated open questions regarding the Cartesian product of the cycles C , m 2 f11; 13; 17g, with paths of arbitrary lengths. Question 2. What is the star chromatic index of C  P for m 2 f11; 13; 17g and n  7? m n As we observed in Theorems 14 and 15, the star chromatic index increases for the Cartesian products of the cycles C and C with paths on at least 7 vertices. We expect this phenomenon 5 7 will repeat also for some cycle C , where m 2 f11; 13; 17g, although, in the case of C and C , m 11 17 we found star 6-edge-colorings of C  P and C  P , which could indicate that 6 colors are 11 8 17 8 sufficient in these cases. To conclude, as in the case of complete graphs, also for the Cartesian products of paths and cycles (and two cycles), it is hard to determine their chromatic index with our current methods. However, it seems that the latter will be easier to resolve as the former. Acknowledgment. PH and MM acknowledge the financial support from the project GA20– 09525S of the Czech Science Foundation. BL was partly supported by the Slovenian Research Agency Program P1–0383 and the project J1–1692. EM and RS were supported by the Slovak Research and Development Agency under the Contract No. APVV–15–0116 and by the Slovak VEGA Grant 1/0368/16. References [1] L’. Bezegová, B. Lužar, M. Mockovčiaková, R. Soták, and R. Škrekovski. Star edge coloring of some classes of graphs. J. Graph Theory, 81:73–82, 2016. 13 [2] C. J. Casselgren, J. B. Granholm, and A. Raspaud. On star edge colorings of bipartite and subcubic graphs. 2019. ArXiv Preprint (https://arxiv.org/pdf/1912.02467.pdf). [3] Z. Dvořák, B. Mohar, and R. Šámal. Star chromatic index. J. Graph Theory, 72:313–326, 2013. [4] B. Grünbaum. Acyclic coloring of planar graphs. Israel J. Math., 14:390–412, 1973. [5] S. Kerdjoudj and A. Raspaud. List star edge coloring of sparse graphs. Discrete Appl. Math., 238:115–125, 2018. [6] H. Lei, Y. Shi, and Z.-X. Song. Star chromatic index of subcubic multigraphs. J. Graph Theory, 88:566–576, 2018. [7] H. Lei, Y. Shi, Z.-X. Song, and T. Wang. Star 5-edge-colorings of subcubic multigraphs. Discrete Math., 341:950–956, 2018. [8] X.-S. Liu and K. Deng. An upper bound on the star chromatic index of graphs with   7. J. Lanzhou Univ. (Nat. Sci.), 44:94–95, 2008. [9] B. Lužar, M. Mockovčiaková, and R. Soták. Note on list star edge-coloring of subcubic graphs. J. Graph Theory, 90:304–310, 2019. [10] M. Mockovčiaková. Distance Constrained Edge Colorings of Graphs. PhD thesis, P. J. Šafárik University, Faculty of Science, Košice, 2013. [11] B. Omoomi and M. V. Dastjerdi. Star Edge Coloring of the Cartesian Product of Graphs. 2018. Preprint (https://arxiv.org/abs/1802.01300). [12] J. J. Sylvester. On Subvariants, i.e. Semi-Invariants to Binary Quantics of an Unlimited Order. Amer. J. Math., 5:79–136, 1882. [13] Y. Wang, W. Wang, and Y. Wang. Edge-partition and star chromatic index. Appl. Math. Comput., 333:480–489, 2018. [14] Y. Wang, Y. Wang, and W. Wang. Star edge-coloring of graphs with maximum degree four. Appl. Math. Comput., 340:268–275, 2019. 14 A Cartesian products of two cycles Figure 1: Cartesian products of C with cycles 6 0 0 6 5 3 4 0 1 2 0 0 3 4 0 3 4 5 1 2 1 2 4 6 4 0 2 0 1 2 0 4 5 3 3 5 3 5 1 2 0 1 4 2 6 1 4 6 0 0 6 5 3 4 (a) A star 6-edge-coloring of C  C (b) A star 7-edge-coloring of C  C 3 3 3 4 4 3 6 4 6 0 1 2 0 4 6 5 0 5 1 2 3 0 4 3 5 4 6 5 2 4 3 0 5 1 2 1 2 0 1 0 1 0 4 1 6 3 0 4 5 4 3 5 4 2 3 5 6 2 4 5 1 2 0 1 2 3 0 6 5 0 1 0 2 5 3 5 4 3 5 1 4 3 4 3 6 4 6 0 1 2 0 4 6 5 (d) A star 7-edge-coloring of C  C including a star 3 7 (c) A star 7-edge-coloring of C  C 3 5 6-edge-coloring of C  C (darker vertices) 3 3 0 1 2 0 1 0 1 5 4 3 5 4 6 5 4 2 4 1 2 0 1 0 3 5 0 5 4 3 5 4 2 1 3 5 2 0 1 2 3 6 0 6 3 5 4 3 5 4 2 4 3 0 0 0 5 1 2 1 1 (e) A star 7-edge-coloring of C  C including a star 6-edge-coloring of C  C (darker vertices) 3 8 3 3 15 Figure 2: Cartesian products of C with cycles 6 0 0 1 1 0 2 4 1 4 0 5 4 3 2 0 2 4 1 0 3 5 2 3 3 6 1 4 3 5 3 5 2 0 0 1 2 0 5 0 5 2 1 4 0 3 5 2 0 3 4 6 4 0 2 1 4 1 5 4 5 0 2 5 2 1 4 0 5 3 2 0 6 1 6 0 5 3 5 3 2 0 3 1 2 5 4 3 3 2 4 1 0 5 3 0 0 0 1 1 0 2 4 1 4 (a) A star 7-edge-coloring of C  C (b) A star 6-edge-coloring of C  C 4 5 4 6 5 4 2 5 3 3 3 0 1 0 2 0 1 0 4 4 3 5 6 4 6 3 2 5 1 3 0 5 3 0 0 4 4 2 3 2 5 1 3 2 5 1 4 5 3 5 0 0 3 6 0 4 2 4 1 6 5 1 4 5 3 3 5 4 2 3 (c) A star 7-edge-coloring of C  C including a star edge-coloring of C  C (darker vertices) 4 7 4 4 5 3 5 3 4 0 4 0 3 0 1 0 2 0 1 3 2 6 0 4 4 3 5 6 2 1 4 5 3 2 5 1 0 1 5 0 6 3 0 0 4 4 3 4 2 1 2 5 1 3 2 5 1 0 3 4 5 3 5 0 0 6 5 6 2 0 4 2 4 1 5 2 3 5 1 4 5 3 5 3 4 0 4 0 3 (d) A star 7-edge-coloring of C  C including a star edge-coloring of C  C (darker vertices) 4 9 4 4 16 Figure 3: A star 6-edge-coloring of C  C combined of star edge-colorings of 4 10 C  C (lighter vertices) and C  C (darker vertices) 4 4 4 6 1 4 1 4 1 0 2 4 1 4 2 0 3 5 2 4 1 0 3 5 2 3 5 2 0 3 5 3 5 2 0 5 2 0 3 5 2 1 4 0 3 5 4 1 4 1 4 0 2 1 4 1 2 0 5 3 2 1 4 0 5 3 2 5 3 2 0 5 3 5 3 2 0 3 2 0 5 3 2 4 1 0 5 3 1 4 1 4 1 0 2 4 1 4 Figure 4: Cartesian products of C with cycles 6 5 6 4 6 5 4 0 1 3 4 1 3 0 1 2 0 3 4 6 3 4 5 1 2 4 1 4 3 6 5 0 1 5 0 6 3 0 2 6 2 4 1 2 3 2 0 3 1 2 6 3 5 0 0 5 0 1 6 1 2 4 0 5 1 3 1 4 2 1 6 4 5 6 3 1 6 6 5 6 3 4 0 3 2 0 6 3 1 3 1 4 1 0 3 4 5 1 4 1 2 5 4 2 2 5 2 5 2 0 3 3 0 6 3 4 0 3 0 1 4 6 1 2 6 2 5 0 6 6 5 6 4 6 5 4 0 1 3 4 1 (b) A star 7-edge-coloring of C  C , including a star 5 7 (a) A star 7-edge-coloring of C  C 5 5 edge-coloring of C  C (darker vertices) 5 3 2 5 3 1 5 6 3 1 3 0 5 0 1 2 0 3 1 2 0 5 1 4 0 6 4 5 2 4 5 5 2 6 6 3 1 0 3 1 3 6 4 1 3 2 0 1 3 2 4 5 0 3 3 4 5 5 4 2 5 1 2 3 1 0 2 0 4 6 2 4 3 0 6 2 5 6 3 1 3 0 5 1 6 4 1 4 2 5 0 4 6 5 6 0 2 5 6 0 0 6 2 2 2 4 3 1 4 0 3 1 4 0 5 1 4 2 3 1 5 6 3 1 3 0 5 (c) A star 7-edge-coloring of C  C 5 11 17 Figure 5: A star 7-edge-coloring of C  C including a star edge-coloring of 7 7 C C (darker vertices in horizontal direction) and a star edge-coloring of C C 3 7 7 3 (darker vertices in vertical direction) 0 1 0 0 4 3 0 3 4 2 3 2 1 2 3 1 6 6 6 5 5 6 3 0 5 1 4 0 2 3 2 3 4 2 3 4 4 5 6 2 5 1 6 1 5 1 4 0 3 0 3 0 6 3 5 6 4 5 2 6 0 1 2 0 2 1 4 3 5 4 3 1 6 3 2 2 6 5 0 2 6 1 6 0 1 4 5 3 1 4 3 4 6 3 0 2 5 6 2 5 2 6 4 5 0 1 0 0 4 3 0 18 Figure 6: A star 7-edge-coloring of C  C including a star edge-coloring of 11 11 C  C (all darker vertices), which furthermore includes a star edge-coloring of 7 11 C  C (the darkest vertices above) 3 11 0 1 2 5 2 5 6 2 4 2 4 2 6 3 4 0 1 4 0 5 6 5 2 1 0 5 6 3 6 3 1 3 4 3 5 4 2 0 4 5 5 0 4 6 2 0 2 3 1 3 1 2 1 0 0 1 1 3 5 4 0 6 4 0 3 5 3 6 3 0 1 2 2 0 5 6 2 1 2 2 1 6 3 5 1 3 0 1 4 0 5 1 3 0 4 3 2 4 2 0 2 3 0 4 2 5 0 6 3 0 6 5 1 6 4 1 3 1 4 5 1 5 3 3 4 3 5 6 5 2 3 6 2 1 4 2 0 5 2 4 0 1 1 3 3 6 0 1 1 3 5 3 6 0 4 3 0 5 1 3 6 0 0 2 5 2 5 6 2 4 2 2 1 6 1 3 0 1 3 0 1 3 0 1 3 2 4 1 3 4 2 3 6 5 4 1 5 0 6 2 6 1 5 4 0 6 1 2 3 5 4 5 0 3 2 1 4 5 0 6 1 2 1 3 6 4 0 5 2 0 4 4 6 0 0 2 1 6 3 1 3 1 5 0 3 6 4 0 5 1 0 6 1 0 1 2 5 2 5 6 2 4 2 4 19 B Cartesian products of paths and cycles Figure 7: Cartesian products of cycles and P 2 0 4 1 0 0 5 3 2 3 5 4 4 1 0 4 2 1 0 5 2 3 3 2 3 1 5 1 0 5 4 3 0 5 1 2 2 1 5 4 3 0 5 4 1 5 1 3 2 3 3 2 5 4 1 2 0 4 1 0 0 (a) A star 6-edge-coloring of C  P 5 6 3 1 5 3 1 5 0 2 4 0 2 5 3 1 5 3 1 2 4 0 2 4 1 5 3 1 5 3 4 0 2 4 0 3 1 5 3 1 5 3 4 0 5 3 5 0 2 4 0 2 5 3 2 1 0 4 3 1 5 3 3 1 1 4 0 2 4 2 4 0 1 4 0 2 5 3 0 0 5 3 2 5 1 4 3 1 4 5 1 2 1 0 4 2 2 5 3 0 2 3 4 5 1 0 3 0 4 0 2 1 3 2 0 4 5 2 3 5 1 4 0 5 1 2 3 1 4 2 1 2 3 3 2 0 5 4 0 2 3 4 0 5 1 3 3 1 5 3 0 4 0 5 1 2 0 2 4 0 2 4 5 3 4 0 3 1 5 3 1 5 3 1 2 1 2 5 4 4 0 2 4 0 0 5 3 4 1 3 1 5 3 1 5 3 4 0 5 3 5 (c) A star 6-edge-coloring of C  P including a 10 6 (b) A star 6-edge-coloring of C  P star edge-coloring of C  P (darker vertices) 7 6 3 6 20 Figure 8: Cartesian products of cycles and P 2 3 0 2 1 0 2 1 2 3 4 2 3 4 2 3 1 5 4 3 5 4 3 0 1 5 0 1 5 0 0 2 1 0 2 1 0 2 4 2 3 4 2 3 4 2 5 4 3 5 4 3 1 1 5 0 1 5 0 1 1 0 2 1 0 2 5 0 3 4 2 3 4 2 3 4 4 3 5 4 1 0 4 5 0 1 5 0 1 5 2 1 0 2 3 3 2 3 1 3 4 2 3 4 2 3 3 5 1 0 5 4 1 0 1 5 0 1 5 4 0 2 4 4 2 0 3 2 2 2 3 4 2 3 0 0 5 0 3 5 1 2 5 4 5 0 1 4 2 1 1 1 2 0 4 4 0 4 3 1 2 3 5 5 3 2 3 4 5 2 3 5 1 5 0 1 2 0 1 0 0 0 1 1 0 2 4 0 2 3 4 4 3 2 5 4 5 2 3 4 5 0 3 0 2 5 0 1 3 0 2 4 0 2 1 1 2 2 1 1 3 2 4 4 2 3 1 3 5 0 3 4 5 5 4 0 3 5 0 1 0 2 4 4 2 0 1 1 2 2 1 1 2 3 4 2 3 0 1 3 5 2 3 0 3 5 4 0 2 5 4 5 2 0 4 4 0 2 3 4 2 3 1 1 3 3 1 4 3 2 1 3 5 1 5 0 2 5 4 0 5 0 5 5 0 2 1 0 2 3 4 4 3 2 1 1 4 2 1 4 3 5 4 5 2 1 5 0 3 5 2 3 0 2 1 0 2 1 0 0 3 2 4 4 2 3 4 1 5 3 1 5 0 (a) A star 6-edge-coloring of C  P 11 8 3 2 0 0 2 3 1 2 1 3 4 1 5 0 4 5 5 2 3 4 2 3 1 4 0 1 5 0 1 5 2 3 4 2 3 4 2 3 (b) A star 6-edge-coloring of C  P including a star 14 8 edge-coloring of C  P (darker vertices) 3 8

Journal

Mathematics – arXiv (Cornell University)

Published: May 6, 2020

References

  • Star edge coloring of some classes of graphs
    Bezegová