arXiv++ Combinatorics

A graph $G$ is $(a,b)$-sparse if every nonempty subgraph $H$ satisfies $e(H) \leq a v(H) - b$. We are interested in the conditions under which an $(a,b)$-sparse graph can be partitioned $E(G) = E(G_1) \cup E(G_2)$ such that for $i \in \{1,2\}$ we have that $G_i$ is $(a_i, b_i)$-sparse. Kuperwasser, Samotij, and Wigderson conjectured that a $(m,0)$-sparse graph can be partitioned into a $(1,1)$-sparse graph and a $(m,2m-1)$-sparse graph. We prove the conjecture in full. The Kohayakawa-Kreuter Conjecture for Families claims that $n^{-1/m_2}$ is the threshold function for the random graph being Ramsey a.a.s. for graph families $\mathcal{H}_1, \ldots \mathcal{H}_r$. Kuperwasser, Samotij, and Wigderson motivated their conjecture by proving that it is sufficient to establish the Kohayakawa-Kreuter Conjecture for Families.

The following measure of sparsity of multigraphs refining the maximum average degree: For $a>0$ and an arbitrary real $b$, a multigraph $H$ is \emph{$(a,b)$-sparse} if it is loopless and for every $A\subseteq V(H)$ with $|A|\geq 2$, the induced subgraph $H[A]$ has at most $a|A|+b$ edges. Forests are exactly $(1,-1)$-sparse multigraphs. It is known that the vertex set of any $(2,-1)$-sparse multigraph can be partitioned into two parts each of which induces a forest. For a given parameter $D$ we study for which pairs $(a,b)$ every $(a,b)$-sparse multigraph $G$ admits a vertex partition $(V_1, V_2)$ of $V(G)$ such that $G[V_1]$ and $G[V_2]$ are forests, and in addition either (i) $Δ(G[V_1])\leq D$ or (ii) every component of $G[V_1]$ has at most $D$ edges. We find exact bounds on $a$ and $b$ for both types of problems (i) and (ii). We also consider problems of type (i) in the class of simple graphs and find exact bounds for all $D\geq 2$.

Borodin and Kostochka proved that for $d_2 \geq 2d_1+2$ and a graph $G$ where every subgraph $H$ satisfies $$ e(H) < \left(2 - \frac{d_2+2}{(d_1+2)(d_2+1)}\right)n(H) + \frac{1}{d_2+1} $$ has a vertex partition $V(G) = V_1 \cup V_2$ such that $G[V_i]$ has maximum degree at most $d_i$ for each $i$. We show that under the same conditions we can additionally conclude that each $G[V_i]$ is a forest.

We are interested in the general question: to what extent are the spectral properties of a graph connected to the distance function? Our motivation is a concrete example of this question that is due to Steinerberger. We provide some negative results via constructions of families of graphs where the spectral properties disagree with the distance function in a manner that answers many of the questions posed by Steinerberger. We also provide a positive result by replacing Steinerberger's set-up with conditions involving graph curvature.

This paper consists of two halves. In the first half of the paper, we consider real-valued functions $f$ whose domain is the vertex set of a graph $G$ and that are Lipschitz with respect to the graph distance. By placing a uniform distribution on the vertex set, we treat $f$ as a random variable. We investigate the link between the isoperimetric function of $G$ and the functions $f$ that have maximum variance or meet the bound established by the subgaussian inequality. We present several results describing the extremal functions, and use those results to resolve: (A) a conjecture by Bobkov, Houdré, and Tetali characterizing the extremal functions of the subgaussian inequality of the odd cycle, and (B) a conjecture by Alon, Boppana, and Spencer on the relationship between maximum variance functions and the isoperimetric function of product graphs. While establishing a discrete analogue of the curved Brunn-Minkowski inequality for the discrete hypercube, Ollivier and Villani suggested several avenues for research. We resolve them in second half of the paper as follows. (1) They propose that a bound on $t$-midpoints can be obtained by repeated application of the bound on midpoints, if the original sets are convex. We construct a specific example where this reasoning fails, and then prove our construction is general by characterizing the convex sets in the discrete hypercube. (2) A second proposed technique to bound $t$-midpoints involves new results in concentration of measure. We follow through on this proposal, with heavy use on results from the first half of the paper. (3) We show that the curvature of the discrete hypercube is not positive or zero.

We approach the problem of counting the number of walks in a digraph from three different perspectives: enumerative combinatorics, linear algebra, and symbolic dynamics.

Many applications in network science have recently been discovered for the "curvature" of a network, but there is no consensus on the definition for this term. A common approach in these applications is to derive from the curvature either a "logical center" of the network or a tree representation of the network (these would only exist when the curvature is negative), but that such structures can be extracted using curvature alone remains largely conjectural. A connection between one type of curvature -- Gromov's hyperbolicity -- and a tree representation has been known for decades, and recently it has also been connected for unweighted graphs to a logical center. We extend the connection between Gromov's hyperbolicity and a logical center to weighted graphs, and we construct counterexamples showing that no other proposed definition for curvature implies the existence of a logical center. We also consider the leading methods to construct a tree representation of the network and the leading methods to measure the quality of the representation, and show that, despite wildly different descriptions, they are asymptotically equivalent. These results resolve several conjectures, including a conjecture of Dourisboure and Gavoille on a $2$-approximation method for calculating tree-length and all of the conjectures from Jonckheere, Lou, Bonahon, and Baryshnikov relating congestion to rotational symmetry.

A star $k$-coloring is a proper $k$-coloring where the union of two color classes induces a star forest. While every planar graph is 4-colorable, not every planar graph is star 4-colorable. One method to produce a star 4-coloring is to partition the vertex set into a 2-independent set and a forest; such a partition is called an I,F-partition. We use a combination of potential functions and discharging to prove that every graph with maximum average degree less than $\frac{5}{2}$ has an I,F-partition, which is sharp and answers a question of Cranston and West [A guide to the discharging method, arXiv:1306.4434]. This result implies that planar graphs of girth at least 10 are star 4-colorable, improving upon previous results of Bu, Cranston, Montassier, Raspaud, and Wang [Star coloring of sparse graphs, J. Graph Theory 62 (2009), 201-219].

The expansion of a graph is typically associated with its spectral properties - testing whether a graph is an expander is usually done using Cheeger's inequality. One can also use multiple eigenvalues in a higher-order Cheeger's inequality to test a deeper set of properties on the graph. However Cheeger's inquality, and the higher-order Cheeger's inequality, can be imprecise tools. Recently Lee, Gharan, and Trevisan constructed the Noisy Hypercube to prove the sharpness of the gap between spectral expansion and edge expansion in the higher-order Cheeger's inequality. We are concerned with the dual problem: using the upper end of the Laplacian spectrum to test a graph's bipartite nature. This has been shown to have several applications, and recently a dual version of Cheeger's inequality and a dual version of the higher-order Cheeger's inequality have been presented. We construct the Bipartite Noisy Hypercube and use it to prove the sharpness of the gap between spectral bipartite expansion and bipartite edge expansion in the dual version of the dual higher-order Cheeger's inequality.

For a graph $G$, let $G^2$ be the graph with the same vertex set as $G$ and $xy \in E(G^2)$ when $x \neq y$ and $d_G(x,y) \leq 2$. Bonamy, Lévêque, and Pinlou conjectured that if $mad (G) < 4 - \frac{2}{c+1}$ and $Δ(G)$ is large, then $χ_\ell(G^2) \leq Δ(G) + c$. We prove that if $c \geq 3$, $mad (G) < 4 - \frac{4}{c+1}$, and $Δ(G)$ is large, then $χ_\ell(G^2) \leq Δ(G) + c$. Dvořák, Krá\soft{l}, Nejedlý, and Škrekovski conjectured that $χ(G^2) \leq Δ(G) +2$ when $Δ(G)$ is large and $G$ is planar with girth at least $5$; our result implies $χ(G^2) \leq Δ(G) +6$.

For a given graph, $G$, let $A$ be the adjacency matrix, $D$ is the diagonal matrix of degrees, $L' = D - A$ is the combinatorial Laplacian, and $L = D^{-1/2}L'D^{-1/2}$ is the normalized Laplacian. Recently, the eigenvectors corresponding to the smallest eigenvalues of $L$ and $L'$ have been of great interest because of their application to community detection, which is a nebulously defined problem that essentially seeks to find a vertex set $S$ such that there are few edges incident with exactly one vertex of $S$. The connection between community detection and the second smallest eigenvalue (and the corresponding eigenvector) is well-known. The $k$ smallest eigenvalues have been used heuristically to find multiple communities in the same graph, and a justification with theoretical rigor for the use of $k \geq 3$ eigenpairs has only been found very recently. The largest eigenpair of $L$ has been used more classically to solve the MAX-CUT problem, which seeks to find a vertex set $S$ that maximizes the number of edges incident with exactly one vertex of $S$. Very recently Trevisan presented a connection between the largest eigenvalue of $L$ and a recursive approach to the MAX-CUT problem that seeks to find a "bipartite community" at each stage. This is related to Kleinberg's HITS algorithm that finds the largest eigenvalue of $A^TA$. We will provide a justification with theoretical rigor for looking at the $k$ largest eigenvalues of $L$ to find multiple bipartite communities in the same graph, and then provide a heuristic algorithm to find strong bipartite communities that is based on the intuition developed by the theoretical methods. Finally, we will present the results of applying our algorithm to various data-mining problems.

A graph $G$ is $k$-{\em critical} if it has chromatic number $k$, but every proper subgraph of $G$ is $(k-1)$--colorable. Let $f_k(n)$ denote the minimum number of edges in an $n$-vertex $k$-critical graph. Recently the authors gave a lower bound, $f_k(n) \geq \left\lceil \frac{(k+1)(k-2)|V(G)|-k(k-3)}{2(k-1)}\right\rceil$, that solves a conjecture by Gallai from 1963 and is sharp for every $n\equiv 1\,({\rm mod }\, k-1)$. It is also sharp for $k=4$ and every $n\geq 6$. In this paper we refine the result by describing all $n$-vertex $k$-critical graphs $G$ with $|E(G)|= \frac{(k+1)(k-2)|V(G)|-k(k-3)}{2(k-1)}$. In particular, this result implies exact values of $f_5(n)$ when $n\geq 7$.

By the Grunbaum-Aksenov Theorem (extending Grotzsch's Theorem) every planar graph with at most three triangles is 3-colorable. However, there are infinitely many planar 4-critical graphs with exactly four triangles. We describe all such graphs. This answers a question of Erdos from 1990.

A recent lower bound on the number of edges in a k-critical n-vertex graph by Kostochka and Yancey yields a half-page proof of the celebrated Grötzsch Theorem that every planar triangle-free graph is 3-colorable. In this paper we use the same bound to give short proofs of other known theorems on 3-coloring of planar graphs, among whose is the Grünbaum-Aksenov Theorem that every planar with at most three triangles is 3-colorable. We also prove the new result that every graph obtained from a triangle-free planar graph by adding a vertex of degree at most four is 3-colorable.

A graph $G$ is $k$-{\em critical} if it has chromatic number $k$, but every proper subgraph of $G$ is $(k-1)$--colorable. Let $f_k(n)$ denote the minimum number of edges in an $n$-vertex $k$-critical graph. In a very recent paper, we gave a lower bound, $f_k(n) \geq F(k,n)$, that is sharp for every $n=1\,({\rm mod}\, k-1)$. It is also sharp for $k=4$ and every $n\geq 6$. In this note, we present a simple proof of the bound for $k=4$. It implies the case $k=4$ of the conjecture by Ore from 1967 that for every $k\geq 4$ and $n\geq k+2$, $f_k(n+k-1)=f(n)+\frac{k-1}{2}(k - \frac{2}{k-1})$. We also show that our result implies a simple short proof of the Gr\" otzsch Theorem that every triangle-free planar graph is 3-colorable.

A graph $G$ is $k$-critical if it has chromatic number $k$, but every proper subgraph of $G$ is $(k-1)$--colorable. Let $f_k(n)$ denote the minimum number of edges in an $n$-vertex $k$-critical graph. We give a lower bound, $f_k(n) \geq F(k,n)$, that is sharp for every $n=1 ({\rm mod} k-1)$. It is also sharp for $k=4$ and every $n\geq 6$. The result improves the classical bounds by Gallai and Dirac and subsequent bounds by Krivelevich and Kostochka and Stiebitz. It establishes the asymptotics of $f_k(n)$ for every fixed $k$. It also proves that the conjecture by Ore from 1967 that for every $k\geq 4$ and $n\geq k+2$, $f_k(n+k-1)=f(n)+\frac{k-1}{2}(k - \frac{2}{k-1})$ holds for each $k\geq 4$ for all but at most $k^3/12$ values of $n$. We give a polynomial-time algorithm for $(k-1)$-coloring a graph $G$ that satisfies $|E(G[W])| < F_k(|W|)$ for all $W \subseteq V(G)$, $|W| \geq k$. We also present some applications of the result.

A \textit{rainbow subgraph} of an edge-colored graph is a subgraph whose edges have distinct colors. The \textit{color degree} of a vertex $v$ is the number of different colors on edges incident to $v$. We show that if $n$ is large enough (namely, $n\geq 4.25k^2$), then each $n$-vertex graph $G$ with minimum color degree at least $k$ contains a rainbow matching of size at least $k$.