Searcharxiv⌕ Search

arXiv subjects

Louis Esperet

Publications and source records attributed to Louis Esperet.

At least 73 records · Page 4Linked to original sources
arXiv2016

Box representations of embedded graphs

A $d$-box is the cartesian product of $d$ intervals of $\mathbb{R}$ and a $d$-box representation of a graph $G$ is a representation of $G$ as the intersection graph of a set of $d$-boxes in $\mathbb{R}^d$. It was proved by Thomassen in 1986 that every planar graph has a 3-box representation. In this paper we prove that every graph embedded in a fixed orientable surface, without short non-contractible cycles, has a 5-box representation. This directly implies that there is a function $f$, such that in every graph of genus $g$, a set of at most $f(g)$ vertices can be removed so that the resulting graph has a 5-box representation. We show that such a function $f$ can be made linear in $g$. Finally, we prove that for any proper minor-closed class $\mathcal{F}$, there is a constant $c(\mathcal{F})$ such that every graph of $\mathcal{F}$ without cycles of length less than $c(\mathcal{F})$ has a 3-box representation, which is best possible.

math.CO↗
arXiv2016

Restricted frame graphs and a conjecture of Scott

Scott proved in 1997 that for any tree $T$, every graph with bounded clique number which does not contain any subdivision of $T$ as an induced subgraph has bounded chromatic number. Scott also conjectured that the same should hold if $T$ is replaced by any graph $H$. Pawlik et al. recently constructed a family of triangle-free intersection graphs of segments in the plane with unbounded chromatic number (thereby disproving an old conjecture of Erdős). This shows that Scott's conjecture is false whenever $H$ is obtained from a non-planar graph by subdividing every edge at least once. It remains interesting to decide which graphs $H$ satisfy Scott's conjecture and which do not. In this paper, we study the construction of Pawlik et al. in more details to extract more counterexamples to Scott's conjecture. For example, we show that Scott's conjecture is false for any graph obtained from $K_4$ by subdividing every edge at least once. We also prove that if $G$ is a 2-connected multigraph with no vertex contained in every cycle of $G$, then any graph obtained from $G$ by subdividing every edge at least twice is a counterexample to Scott's conjecture.

math.CO↗
arXiv2015

Graphs with no induced five-vertex path or antipath

We prove that a graph $G$ contains no induced $5$-vertex path and no induced complement of a $5$-vertex path if and only if $G$ is obtained from $5$-cycles and split graphs by repeatedly applying the following operations: substitution, split unification, and split unification in the complement, where split unification is a new class-preserving operation introduced here.

math.CO↗
arXiv2015

Islands in graphs on surfaces

An island in a graph is a set $X$ of vertices, such that each element of $X$ has few neighbors outside $X$. In this paper, we prove several bounds on the size of islands in large graphs embeddable on fixed surfaces. As direct consequences of our results, we obtain that: (1) Every graph of genus $g$ can be colored from lists of size 5, in such a way that each monochromatic component has size $O(g)$. Moreover all but $O(g)$ vertices lie in monochromatic components of size at most 3. (2) Every triangle-free graph of genus $g$ can be colored from lists of size 3, in such a way that each monochromatic component has size $O(g)$. Moreover all but $O(g)$ vertices lie in monochromatic components of size at most 10. (3) Every graph of girth at least 6 and genus $g$ can be colored from lists of size 2, in such a way that each monochromatic component has size $O(g)$. Moreover all but $O(g)$ vertices lie in monochromatic components of size at most 16. While (2) is optimal up to the size of the components, we conjecture that the size of the lists can be decreased to 4 in (1), and the girth can be decreased to 5 in (3). We also study the complexity of minimizing the size of monochromatic components in 2-colorings of planar graphs.

math.CO↗
arXiv2015

The structure of graphs with Circular flow number 5 or more, and the complexity of their recognition problem

For some time the Petersen graph has been the only known Snark with circular flow number $5$ (or more, as long as the assertion of Tutte's $5$-flow Conjecture is in doubt). Although infinitely many such snarks were presented eight years ago by Macajova and Raspaud, the variety of known methods to construct them and the structure of the obtained graphs were still rather limited. We start this article with an analysis of sets of flow values, which can be transferred through flow networks with the flow on each edge restricted to the open interval $(1,4)$ modulo $5$. All these sets are symmetric unions of open integer intervals in the ring $\mathbb{R}/5\mathbb{Z}$. We use the results to design an arsenal of methods for constructing snarks $S$ with circular flow number $ϕ_c(S)\ge 5$. As one indication to the diversity and density of the obtained family of graphs, we show that it is sufficiently rich so that the corresponding recognition problem is NP-complete.

math.CO↗
arXiv2015

Boxicity and topological invariants

The boxicity of a graph $G=(V,E)$ is the smallest integer $k$ for which there exist $k$ interval graphs $G_i=(V,E_i)$, $1 \le i \le k$, such that $E=E_1 \cap \cdots \cap E_k$. In the first part of this note, we prove that every graph on $m$ edges has boxicity $O(\sqrt{m \log m})$, which is asymptotically best possible. We use this result to study the connection between the boxicity of graphs and their Colin de Verdière invariant, which share many similarities. Known results concerning the two parameters suggest that for any graph $G$, the boxicity of $G$ is at most the Colin de Verdière invariant of $G$, denoted by $μ(G)$. We observe that every graph $G$ has boxicity $O(μ(G)^4(\log μ(G))^2)$, while there are graphs $G$ with boxicity $Ω(μ(G)\sqrt{\log μ(G)})$. In the second part of this note, we focus on graphs embeddable on a surface of Euler genus $g$. We prove that these graphs have boxicity $O(\sqrt{g}\log g)$, while some of these graphs have boxicity $Ω(\sqrt{g \log g})$. This improves the previously best known upper and lower bounds. These results directly imply a nearly optimal bound on the dimension of the adjacency poset of graphs on surfaces.

math.CO↗
arXiv2015

Equitable partition of graphs into induced forests

An equitable partition of a graph $G$ is a partition of the vertex-set of $G$ such that the sizes of any two parts differ by at most one. We show that every graph with an acyclic coloring with at most $k$ colors can be equitably partitioned into $k-1$ induced forests. We also prove that for any integers $d\ge 1$ and $k\ge 3^{d-1}$, any $d$-degenerate graph can be equitably partitioned into $k$ induced forests. Each of these results implies the existence of a constant $c$ such that for any $k \ge c$, any planar graph has an equitable partition into $k$ induced forests. This was conjectured by Wu, Zhang, and Li in 2013.

math.CO↗
arXiv2015

On the maximum fraction of edges covered by t perfect matchings in a cubic bridgeless graph

A conjecture of Berge and Fulkerson (1971) states that every cubic bridgeless graph contains 6 perfect matchings covering each edge precisely twice, which easily implies that every cubic bridgeless graph has three perfect matchings with empty intersection (this weaker statement was conjectured by Fan and Raspaud in 1994). Let $m_t$ be the supremum of all reals $α\le 1$ such that for every cubic bridgeless graph $G$, there exist $t$ perfect matchings of $G$ covering a fraction of at least $α$ of the edges of $G$. It is known that the Berge-Fulkerson conjecture is equivalent to the statement that $m_5=1$, and implies that $m_4=\tfrac{14}{15}$ and $m_3=\tfrac45$. In the first part of this paper, we show that $m_4=\tfrac{14}{15}$ implies $m_3=\tfrac45$, and $m_3=\tfrac45$ implies the Fan-Raspaud conjecture, strengthening a recent result of Tang, Zhang, and Zhu. In the second part of the paper, we prove that for any $2\le t \le 4$ and for any real $τ$ lying in some appropriate interval, deciding whether a fraction of more than (resp. at least) $τ$ of the edges of a given cubic bridgeless graph can be covered by $t$ perfect matching is an NP-complete problem. This resolves a conjecture of Tang, Zhang, and Zhu.

math.CO↗
arXiv2014

Coloring planar graphs with three colors and no large monochromatic components

We prove the existence of a function $f :\mathbb{N} \to \mathbb{N}$ such that the vertices of every planar graph with maximum degree $Δ$ can be 3-colored in such a way that each monochromatic component has at most $f(Δ)$ vertices. This is best possible (the number of colors cannot be reduced and the dependence on the maximum degree cannot be avoided) and answers a question raised by Kleinberg, Motwani, Raghavan, and Venkatasubramanian in 1997. Our result extends to graphs of bounded genus.

math.CO↗
arXiv2013

On cubic bridgeless graphs whose edge-set cannot be covered by four perfect matchings

The problem of establishing the number of perfect matchings necessary to cover the edge-set of a cubic bridgeless graph is strictly related to a famous conjecture of Berge and Fulkerson. In this paper we prove that deciding whether this number is at most 4 for a given cubic bridgeless graph is NP-complete. We also construct an infinite family $\cal F$ of snarks (cyclically 4-edge-connected cubic graphs of girth at least five and chromatic index four) whose edge-set cannot be covered by 4 perfect matchings. Only two such graphs were known. It turns out that the family $\cal F$ also has interesting properties with respect to the shortest cycle cover problem. The shortest cycle cover of any cubic bridgeless graph with $m$ edges has length at least $\tfrac43m$, and we show that this inequality is strict for graphs of $\cal F$. We also construct the first known snark with no cycle cover of length less than $\tfrac43m+2$.

math.CO↗
arXiv2013

Distance-two coloring of sparse graphs

Consider a graph $G = (V, E)$ and, for each vertex $v \in V$, a subset $Σ(v)$ of neighbors of $v$. A $Σ$-coloring is a coloring of the elements of $V$ so that vertices appearing together in some $Σ(v)$ receive pairwise distinct colors. An obvious lower bound for the minimum number of colors in such a coloring is the maximum size of a set $Σ(v)$, denoted by $ρ(Σ)$. In this paper we study graph classes $F$ for which there is a function $f$, such that for any graph $G \in F$ and any $Σ$, there is a $Σ$-coloring using at most $f(ρ(Σ))$ colors. It is proved that if such a function exists for a class $F$, then $f$ can be taken to be a linear function. It is also shown that such classes are precisely the classes having bounded star chromatic number. We also investigate the list version and the clique version of this problem, and relate the existence of functions bounding those parameters to the recently introduced concepts of classes of bounded expansion and nowhere-dense classes.

math.CO↗
arXiv2013

Acyclic edge-coloring using entropy compression

An edge-coloring of a graph G is acyclic if it is a proper edge-coloring of G and every cycle contains at least three colors. We prove that every graph with maximum degree Delta has an acyclic edge-coloring with at most 4 Delta - 4 colors, improving the previous bound of 9.62 (Delta - 1). Our bound results from the analysis of a very simple randomised procedure using the so-called entropy compression method. We show that the expected running time of the procedure is O(mn Delta^2 log Delta), where n and m are the number of vertices and edges of G. Such a randomised procedure running in expected polynomial time was only known to exist in the case where at least 16 Delta colors were available. Our aim here is to make a pedagogic tutorial on how to use these ideas to analyse a broad range of graph coloring problems. As an application, also show that every graph with maximum degree Delta has a star coloring with 2 sqrt(2) Delta^{3/2} + Delta colors.

math.CO↗
arXiv2012

Fire Containment in Planar Graphs

In a graph $G$, a fire starts at some vertex. At every time step, firefighters can protect up to $k$ vertices, and then the fire spreads to all unprotected neighbours. The $k$-surviving rate $ρ_k(G)$ of $G$ is the expectation of the proportion of vertices that can be saved from the fire, if the starting vertex of the fire is chosen uniformly at random. For a given class of graphs $\cG$ we are interested in the minimum value $k$ such that $ρ_k(G)\geε$ for some constant $ε>0$ and all $G\in\cG$ i.e., such that linearly many vertices are expected to be saved in every graph from $\cG$). In this note, we prove that for planar graphs this minimum value is at most 4, and that it is precisely 2 for triangle-free planar graphs.

math.CO↗
arXiv2012

A Unified Approach to Distance-Two Colouring of Graphs on Surfaces

In this paper we introduce the notion of $Σ$-colouring of a graph $G$: For given subsets $Σ(v)$ of neighbours of $v$, for every $v\in V(G)$, this is a proper colouring of the vertices of $G$ such that, in addition, vertices that appear together in some $Σ(v)$ receive different colours. This concept generalises the notion of colouring the square of graphs and of cyclic colouring of graphs embedded in a surface. We prove a general result for graphs embeddable in a fixed surface, which implies asymptotic versions of Wegner's and Borodin's Conjecture on the planar version of these two colourings. Using a recent approach of Havet et al., we reduce the problem to edge-colouring of multigraphs, and then use Kahn's result that the list chromatic index is close to the fractional chromatic index. Our results are based on a strong structural lemma for graphs embeddable in a fixed surface, which also implies that the size of a clique in the square of a graph of maximum degree $Δ$ embeddable in some fixed surface is at most $\frac32\,Δ$ plus a constant.

math.CO↗
arXiv2012

Locally identifying coloring of graphs

We introduce the notion of locally identifying coloring of a graph. A proper vertex-coloring c of a graph G is said to be locally identifying, if for any adjacent vertices u and v with distinct closed neighborhood, the sets of colors that appear in the closed neighborhood of u and v are distinct. Let $χ_{lid}(G)$ be the minimum number of colors used in a locally identifying vertex-coloring of G. In this paper, we give several bounds on $χ_{lid}$ for different families of graphs (planar graphs, some subclasses of perfect graphs, graphs with bounded maximum degree) and prove that deciding whether $χ_{lid}(G)=3$ for a subcubic bipartite graph $G$ with large girth is an NP-complete problem.

cs.DM↗
arXiv2012

Boxicity of graphs on surfaces

The boxicity of a graph $G=(V,E)$ is the least integer $k$ for which there exist $k$ interval graphs $G_i=(V,E_i)$, $1 \le i \le k$, such that $E=E_1 \cap ... \cap E_k$. Scheinerman proved in 1984 that outerplanar graphs have boxicity at most two and Thomassen proved in 1986 that planar graphs have boxicity at most three. In this note we prove that the boxicity of toroidal graphs is at most 7, and that the boxicity of graphs embeddable in a surface $Σ$ of genus $g$ is at most $5g+3$. This result yields improved bounds on the dimension of the adjacency poset of graphs on surfaces.

math.CO↗
arXiv2011

Exponentially many perfect matchings in cubic graphs

We show that every cubic bridgeless graph G has at least 2^(|V(G)|/3656) perfect matchings. This confirms an old conjecture of Lovasz and Plummer. This version of the paper uses a different definition of a burl from the journal version of the paper and a different proof of Lemma 18 is given. This simplifies the exposition of our arguments throughout the whole paper.

math.CO↗