Abstract
We prove that for all integers \(\Delta ,r \ge 2\), there is a constant \(C = C(\Delta ,r) >0\) such that the following is true for every sequence \({\mathcal {F}}= \{F_1, F_2, \ldots \}\) of graphs with \(v(F_n) = n\) and \(\Delta (F_n) \le \Delta \), for each \(n \in {\mathbb {N}}\). In every r-edge-coloured \(K_n\), there is a collection of at most C monochromatic copies from \({\mathcal {F}}\) whose vertex-sets partition \(V(K_n)\). This makes progress on a conjecture of Grinshpun and Sárközy.
Similar content being viewed by others
Notes
The k-th power of a graph H is the graph obtained from H by adding an edge between any two vertices at distance at most k.
A k-clique is transversal in \((V_1,\ldots ,V_k)\) if it contains one vertex in each one of the sets \(V_1,\ldots ,V_k\).
The second part of the theorem is not explicitly stated in [17] but follows readily from the blow-up lemma and the Hajnal-Szemerédi theorem.
Here, we denote by [i, j] the set of integers z with \(i \le z \le j\).
Note that the constant from Proposition 4.2 is smaller than \(C'\).
We shall identify the cylinders with their vertex-set.
That is, \(F_i \subseteq F_{i+1}\), for every \(i \in {\mathbb {N}}\).
A graph G is d-degenerate if there is an ordering of its vertices so that every \(v \in V(G)\) is adjacent to at most d vertices which come before v.
A graph G is called a-arrangeable for some \(a \in {\mathbb {N}}\) if its vertices can be ordered in such a way that for every \(v \in V(G)\), there are at most a vertices to the left of v that have some common neighbour with v to the right of v.
References
Allen, P.: Covering two-edge-coloured complete graphs with two disjoint monochromatic cycles. Combin. Probab. Comput. 17(4), 471–486 (2008)
Ayel, J.: Sur l’existence de deux cycles supplémentaires unicolores, disjoints et de couleurs différentes dans un graphe complet bicolore, Theses, Université Joseph-Fourier — Grenoble I, (1979)
Bessy, S., Thomassé, S.: Partitioning a graph into a cycle and an anticycle, a proof of Lehel’s conjecture. J. Combin. Theory Ser. B 100(2), 176–180 (2010)
Böttcher, J., Kohayakawa, Y., Taraz, A., Würfl, A.: An extension of the blow-up lemma to arrangeable graphs. SIAM J. Discrete Math. 29(2), 962–1001 (2015)
Bustamante, S., Corsten, J., Frankl, N., Pokrovskiy, A., Skokan, J.: Partitioning hypergraphs into few monochromatic tight cycles. SIAM J. Discrete Math. 34(2), 1460–1471 (2020)
Chen, G.T., Schelp, R.H.: Graphs with linearly bounded Ramsey numbers. J. Combin. Theory Ser. B 57(1), 138–149 (1993)
Conlon, D., Fox, J.: Bounds for graph regularity and removal lemmas. Geom. Funct. Anal. 22(5), 1191–1256 (2012)
Conlon, D., Fox, J., Sudakov, B.: On two problems in graph Ramsey theory. Combinatorica 32(5), 513–535 (2012)
Duke, R.A., Lefmann, H., Rödl, V.: A fast approximation algorithm for computing the frequencies of subgraphs in a given graph. SIAM J. Comput. 24(3), 598–620 (1995)
Elekes, M., Soukup, D.T., Soukup, L., Szentmiklóssy, Z.: Decompositions of edge-colored infinite complete graphs into monochromatic paths. Discrete Math. 340(8), 2053–2069 (2017)
Erdős, P., Gyárfás, A., Pyber, L.: Vertex coverings by monochromatic cycles and trees. J. Combin. Theory Ser. B 51(1), 90–95 (1991)
Erdös, P., Szekeres, G.: A combinatorial problem in geometry. Compos. Math. 2, 463–470 (1935)
Fox, J., Sudakov, B.: Density theorems for bipartite graphs and related Ramsey-type results. Combinatorica 29(2), 153–196 (2009)
Fox, J., Lovász, L.M.: A tight lower bound for Szemerédi’s regularity lemma. Combinatorica 37(5), 911–951 (2017)
Frieze, A.M., Kannan, R.: The regularity lemma and approximation schemes for dense problems. In: Proceedings of 37th Conference on Foundations of Computer Science, IEEE, pp. 12–20 (1996)
Graham, R.L., Rödl, V., Ruciński, A.: On graphs with linear Ramsey numbers. J. Graph Theory 35(3), 176–192 (2000)
Grinshpun, A., Sárközy, G.N.: Monochromatic bounded degree subgraph partitions. Discrete Math. 339(1), 46–53 (2016)
Gyárfás, A., Ruszinkó, M., Sárközy, G.N., Szemerédi, E.: An improved bound for the monochromatic cycle partition number. J. Combin. Theory Ser. B 96(6), 855–873 (2006)
Hajnal, A., Szemerédi, E.: Proof of a conjecture of P. Erdős, Combinatorial theory and its applications, II (Proc. Colloq., Balatonfüred, 1969), pp. 601–623 (1970)
Komlós, J., Sárközy, G.N., Szemerédi, E.: Blow-up lemma. Combinatorica 17(1), 109–123 (1997)
Komlós, J., Sárközy, G.N., Szemerédi, E.: An algorithmic version of the blow-up lemma. Random Struct. Algorithms 12(3), 297–312 (1998)
Lee, C.: Ramsey numbers of degenerate graphs. Ann. Math. 185(3), 791–829 (2017)
Łuczak, T., Rödl, V., Szemerédi, E.: Partitioning two-coloured complete graphs into two monochromatic cycles. Combin. Probab. Comput. 7(4), 423–436 (1998)
Pokrovskiy, A.: Partitioning edge-coloured complete graphs into monochromatic cycles and paths. J. Combin. Theory Ser. B 106, 70–97 (2014)
Ramsey, F.P.: On a problem of formal logic. Proc. Lond. Math. Soc. 30(4), 264–286 (1929)
Sárközy, G.N.: A quantitative version of the Blow-up lemma, arxiv:1405.7302 (2014)
Sárközy, G.N.: Monochromatic cycle power partitions. Discrete Math. 340(2), 72–80 (2017)
Acknowledgements
The authors would like to thank the organisers of the workshop Extremal and Structural Combinatorics, held at IMPA in Rio de Janeiro, where this work began, and thank Rob Morris for reading a previous version of this paper.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Jan Corsten Supported by an LSE studentship. Walner Mendonça Partially supported by CAPES (88882.332408/2010-01) and FAPESP (2023/07695-6).
Appendix
Appendix
In this appendix, we shall prove the lemmas stated in Sect. 3 for which we could not find a proof in the literature. Their proofs however are standard and not difficult.
Proof of Lemma 3.2
Let \(U_i = (V_i \setminus X_i) \cup Y_i\) for \( i \in \{1,2\}\). We will show that \( (U_1,U_2) \) is \((8\varepsilon ,d - 8\varepsilon ,\delta /2)\)-super-regular. Let now \(Z_i \subseteq U_i\) with \(|Z_i| \ge 8 \varepsilon |U_i|\), and let \(Z_i' = Z_i {\setminus } Y_i\) and \(Z_i'' = Z_i \cap Y_i\) for \(i \in \{1,2\}\). Note that we have
for both \(i \in \{1,2\}\). We therefore have
and thus
On the other hand, we have
and hence \(d(Z_1,Z_2) = d(Z_1',Z_2') \pm 2\varepsilon \). Furthermore, by \(\varepsilon \)-regularity of \((V_1,V_2)\), we have \(d(Z_1',Z_2') = d(V_1, V_2) \pm \varepsilon \) and we conclude
This holds in particular for \(Z_1 = U_1\) and \(Z_2 = U_2\) and therefore the pair \((U_1,U_2)\) is \((8\varepsilon ,d-8\varepsilon ,0)\)-super-regular. Let \(u_1 \in U_1\) now. By assumption, we have \(\deg (u_1,V_2) \ge \delta |V_2|\) and therefore
A similar statement is true for every \(u_2 \in U_2\) finishing the proof. \(\square \)
The following consequence of the slicing lemma will be useful when we prove Lemmas 3.4,3.5.
Lemma A.1
Let k be a positive integer and \(d,\varepsilon >0\) with \(\varepsilon \le 1/(2k)\). If \(Z = (V_1, \ldots , V_k)\) is an \(\varepsilon \)-regular k-cylinder and \(d(V_i,V_j) \ge d\) for all \(1 \le i < j \le k\), then there is some \( \gamma \le k\varepsilon \) and sets \({{\tilde{V}}}_1 \subseteq V_1, \ldots , {{\tilde{V}}}_k \subseteq V_k\) with \(|{\tilde{V}}_i| = \lceil (1-\gamma ) |V_i| \rceil \) for all \( i \in [k]\) so that the k-cylinder \(\tilde{Z} = ({{\tilde{V}}}_1, \ldots , {{\tilde{V}}}_k)\) is \((2\varepsilon ,d - k\varepsilon )\)-super-regular.
Proof
For \(i \not = j \in [k]\), let \(A_{i,j}:= \{v \in V_i: \deg (v,V_j) < (d-\varepsilon )|V_j|\}\). By definition of \(\varepsilon \)-regularity, we have \(\left| A_{i,j} \right| < \varepsilon |V_i|\) for every \(i \not = j \in [k]\). For each \(i \in [k]\), let \(A_i=\bigcup _{j \in [k]\setminus \{i\}} A_{i,j}\). Clearly \(|A_i| < (k-1)\varepsilon |V_i|\) for every \(i \in [k]\), so we can add arbitrary vertices from \(V_i \setminus A_i\) to \(A_i\) until \(|A_i| = \lfloor (k-1) \varepsilon |V_i| \rfloor \) for every \(i \in [k]\). Let now \( {{\tilde{V}}}_i = V_i {\setminus } {\tilde{A}}_i\) for every \(i \in [k]\) and let \({{\tilde{Z}}} = ({{\tilde{V}}}_1,\ldots ,{{\tilde{V}}}_k)\). Observe that \(|{\tilde{V}}_i| = \lceil (1 - \gamma )|V_i| \rceil \) for all \(i \in [k]\), where \(\gamma = (k-1)\varepsilon \). It follows from Lemma 3.1 and definition of \(A_i\) that \({{\tilde{Z}}}\) is \((2\varepsilon , d-\varepsilon , d-k\varepsilon )\)-super-regular. \(\square \)
Given k disjoint sets \(V_1,\ldots ,V_k\), we call a cylinder \((U_1, \ldots , U_k)\) relatively balanced (w.r.t. \((V_1, \ldots , V_k)\)) if there exists some \(\gamma > 0 \) so that \(U_i \subseteq V_i\) with \(|U_i| = \lfloor \gamma |V_i| \rfloor \) for every \(i \in [k]\). We say that a partition \({\mathcal {K}}\) of \(V_1\times \cdots \times V_k\) is cylindrical if each partition class is of the form \(W_1 \times \cdots \times W_k\) (which we associate with the k-cylinder \(Z=(W_1, \ldots , W_k)\)) with \(W_j \subseteq V_j\) for every \(j \in [k]\). Finally, we say that \({\mathcal {K}}= \{Z_1, \ldots , Z_N\}\) is \(\varepsilon \)-regular if
-
(i)
\({\mathcal {K}}\) is a cylindrical partition of \(V_1\times \cdots \times V_k\),
-
(ii)
each \(Z_i\), \(i \in [k]\), is a relatively balanced w.r.t. \((V_1,\ldots ,V_k)\), and
-
(iii)
all but \(\varepsilon |V_1|\cdots |V_k|\) of the k-tuples \((v_1,\ldots ,v_k) \in V_1 \times \cdots \times V_k\) are in \(\varepsilon \)-regular cylinders.
For technical reasons, we will allow some of the sets \(V_1,\ldots ,V_k\) to be empty. In this case \((A,\emptyset )\) is considered \(\varepsilon \)-regular for every set A and \(\varepsilon >0\). If G is an r-edge-coloured graph and \(i \in [r]\), we say that a cylinder \({\mathcal {K}}\) is \(\varepsilon \)-regular in colour i if is \(\varepsilon \)-regular in \(G_i\) (the graph on V(G) with all edges of colour i).
In [7], Conlon and Fox used the weak regularity lemma of Duke, Lefmann and Rödl [9] to find a reasonably large balanced k-cylinder in a k-partite graph. In order to prove a coloured version of Conlon and Fox’s result, we will need the following coloured version of the weak regularity lemma of Duke, Lefmann and Rödl. Note that, like the weak regularity lemma of Frieze and Kannan [15], we get an exponential bound on the number of cylinders, in contrast to the much worse tower-type bound required by Szemerédi’s regularity lemma (see [14]).
Theorem A.2
(Duke-Lefmann-Rödl [9]) Let \( 0< \varepsilon < 1/2\), \(k, r \in {\mathbb {N}}\) and let \( \beta = \varepsilon ^{rk^2 \varepsilon ^{-5}}\). Let G be an r-edge-coloured k-partite graph with parts \(V_1, \ldots ,V_k\). Then there exist some \(N \le \beta ^{-k}\), sets \(R_1 \subseteq V_1, \ldots , R_k \subseteq V_k\) with \(\left| R_i \right| \le \beta ^{-1}\) and a partition \({\mathcal {K}}= \{Z_1, \ldots , Z_N\}\) of \((V_1 \setminus R_1) \times \cdots \times (V_k \setminus R_k)\) so that \({\mathcal {K}}\) is \(\varepsilon \)-regular in every colour and \(V_i(Z_j) \ge \lfloor \beta |V_i| \rfloor \) for every \(i \in [k]\) and \(j \in [N]\).
Although the original statement of Duke, Lefmann and Rödl [9, Proposition 2.1] does not involve the colouring and assume that sets \(V_1,\ldots ,V_k\) have the same size, their proof can be easily adapted to prove Theorem A.2.
We are now ready to prove Lemmas 3.4,3.5.
Proof of Lemma 3.4
Let \(k,r \ge 2\), \(0< \varepsilon < 1/(rk)\) and \( \gamma = \varepsilon ^{r^{8rk}\varepsilon ^{-5}}\). Let \(n \ge 1/\gamma \) and suppose we are given an r-edge coloured \(K_n\). Let \({\tilde{k}} = r^{rk}\) and let \(V_1, \ldots , V_{{\tilde{k}}} \subseteq [n]\) be disjoint sets of size \(\lfloor n/{\tilde{k}} \rfloor \) and let G be the \({\tilde{k}}\)-partite subgraph of \(K_n\) induced by \(V_1, \ldots , V_{{\tilde{k}}}\) (inheriting the colouring). Let \(\tilde{\varepsilon } = \varepsilon /2\) and \(\beta = {\tilde{\varepsilon }}^{r^{2rk+1} {\tilde{\varepsilon }}^{-5}}\). We apply Theorem A.2 to get some \(N \le \beta ^{-{{\tilde{k}}}}\), sets \(R_1 \subseteq V_1,\ldots , R_{{{\tilde{k}}}} \subseteq V_{{{\tilde{k}}}}\) each of which of size at most \(\beta ^{-1}\) and a partition \({\mathcal {K}}= \{Z_1, \ldots ,Z_N\}\) of \((V_1 {\setminus } R_i) \times \cdots \times (V_{{{\tilde{k}}}} {\setminus } R_{{{\tilde{k}}}})\) which is \(\tilde{\varepsilon }\)-regular in every colour, and with \(V_i(Z_j) \ge \lfloor \beta |V_i| \rfloor \ge 2 \gamma n\) for every \(i \in [{{\tilde{k}}}]\) and \(j \in [N]\). Note that one of the cylinders (say \(Z_1\)) must be \(\tilde{\varepsilon }\)-regular in every colour and, since \((V_1, \ldots , V_k)\) is balanced, so is \(Z_1\). We consider now the complete graph with vertex-set \(\{V_1(Z_1), \ldots , V_{{\tilde{k}}}(Z_1)\}\) and colour every edge \(V_i(Z_1)V_j(Z_1)\), \(1 \le i < j \le {\tilde{k}}\), with a colour \(c \in [r]\) so that the density of the pair \((V_i(Z_1),V_j(Z_1))\) in colour c is at least 1/r. By Ramsey’s theorem [12, 25], there is a colour, say 1, and k parts (say \(V_1(Z_1), \ldots , V_k(Z_1)\)) so that the cylinder \((V_1(Z_1), \ldots , V_k(Z_1))\) is \((\tilde{\varepsilon },1/r,0)\)-super-regular in colour 1. By Lemma A.1, there is an \((\varepsilon ,1/(2r))\)-super-regular balanced subcylinder \({{\tilde{Z}}}_1\) with parts of size at least \(\gamma n\). \(\square \)
Proof of Lemma 3.5
Let \(k \ge 2\), and let \(d,\varepsilon >0\) with \(2k\varepsilon \le d \le 1\). Let \( \gamma = \varepsilon ^{k^2\varepsilon ^{-12}}\) and let G be a k-partite graph with parts \(V_1, \ldots , V_k\). Let \(\tilde{\varepsilon } = \varepsilon /4\) and \(\beta = \tilde{\varepsilon }^{k^2 \tilde{\varepsilon }^{-5}}\). We may assume that \(|V_i| \ge 1/\gamma \) for every \(i \in [k]\) (otherwise we set \(U_i:= \emptyset \) for all \(i \in [k]\) with \(|V_i| < 1/\gamma \)). In particular, we have \(|V_i| \ge k/({\tilde{\varepsilon }} \beta )\) for all \(i \in [k]\).
We apply Theorem A.2 (with \(r = 1\)) to get some \(N \le \beta ^{-k}\), sets \(R_1 \subseteq V_1,\ldots , R_k \subseteq V_k\), each of which of size at most \(\beta ^{-1}\), and an \(\tilde{\varepsilon }\)-regular partition \({\mathcal {K}}= \{Z_1, \ldots ,Z_N\}\) of \((V_1 {\setminus } R_1) \times \cdots \times (V_k {\setminus } R_k)\) with \(V_i(Z_j) \ge \lfloor \beta |V_i| \rfloor \) for every \(i \in [k]\) and \(j \in [N]\).
Note that the number of cliques of size k incident to \( R = R_1 \cup \ldots \cup R_k\) is at most
Furthermore, since \({\mathcal {K}}\) is \(\tilde{\varepsilon }\)-regular, there are at most \(\tilde{\varepsilon } |V_1| \cdots |V_k|\) cliques of size k in G that belong to a cylinder of \({\mathcal {K}}\) that is not \(\varepsilon \)-regular. Suppose that each cylinder \(Z \in {\mathcal {K}}\) has at most \((d-2\tilde{\varepsilon })|V_1(Z)| \cdots |V_k(Z)|\) cliques of size k. Then the number of k-cliques in G is at most
which contradicts our hypothesis over G. Therefore, there is a cylinder \({{\tilde{Z}}}\) in \({\mathcal {K}}\) that contains at least \((d-2\tilde{\varepsilon })|V_1({{\tilde{Z}}})| \cdots |V_k({{\tilde{Z}}})|\) cliques of size k. In particular, \({{\tilde{Z}}}\) is \((\tilde{\varepsilon },d-2{\tilde{\varepsilon }},0)\)-super-regular and relatively balanced with parts of size at least \(\lfloor \beta |V_i| \rfloor \). Finally, we apply Lemma A.1 (and possibly delete a single vertex from some parts) to get a relatively balanced \((\varepsilon ,d-(k+2)\tilde{\varepsilon })\)-super-regular k-cylinder Z with parts of size at least \(\tfrac{\beta }{2} |V_i| \ge \gamma |V_i|\). This completes the proof since \((k+2)\tilde{\varepsilon } \le k\varepsilon \le d/2\). \(\square \)
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Corsten, J., Mendonça, W. Tiling Edge-Coloured Graphs with Few Monochromatic Bounded-Degree Graphs. Combinatorica 44, 311–335 (2024). https://doi.org/10.1007/s00493-023-00072-1
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00493-023-00072-1