Search

We consider the asymptotic minimum density $f(s,k)$ of monotone $k$-subwords of words over a totally ordered alphabet of size $s$. The unrestricted alphabet case, $f(\infty,k)$, is well-studied, known for $f(\infty,3)$ and $f(\infty,4)$, and, in particular, conjectured to be rational for all $k$. Here we determine $f(2,k)$ for all $k$ and determine $f(3,3)$, which is already irrational. We describe an explicit construction for all $s$ which is conjectured to yield $f(s,3)$. Using our construction and flag algebra, we determine $f(4,3),f(5,3),f(6,3)$ up to $10^{-3}$ yet argue that flag algebra, regardless of computational power, cannot determine $f(5,3)$ precisely. Finally, we prove that for every fixed $k \ge 3$, the gap between $f(s,k)$ and $f(\infty,k)$ is $\Theta(\frac{1}{s})$.

An edge-colored rooted directed tree (aka arborescence) is path-monochromatic if every path in it is monochromatic. Let $k,\ell$ be positive integers. For a tournament $T$, let $f_T(k)$ be the largest integer such that every $k$-edge coloring of $T$ has a path-monochromatic subtree with at least $f_T(k)$ vertices and let $f_T(k,\ell)$ be the restriction to subtrees of depth at most $\ell$. It was proved by Landau that $f_T(1,2)=n$ and proved by Sands et al. that $f_T(2)=n$ where $|V(T)|=n$. Here we consider $f_T(k)$ and $f_T(k,\ell)$ in more generality, determine their extremal values in most cases, and in fact in all cases assuming the Caccetta-H\"aggkvist Conjecture. We also study the typical value of $f_T(k)$ and $f_T(k,\ell)$, i.e., when $T$ is a random tournament.

For integers $k,n$ with $1 \le k \le n/2$, let $f(k,n)$ be the smallest integer $t$ such that every $t$-connected $n$-vertex graph has a spanning bipartite $k$-connected subgraph. A conjecture of Thomassen asserts that $f(k,n)$ is upper bounded by some function of $k$. The best upper bound for $f(k,n)$ is by Delcourt and Ferber who proved that $f(k,n) \le 10^{10}k^3 \log n$. Here it is proved that $f(k,n) \le 22k^2 \log n$. For larger $k$, stronger bounds hold. In the linear regime, it is proved that for any $0 < c < \frac{1}{2}$ and all sufficiently large $n$, if $k=\lfloor cn \rfloor$, then $f(k, n) \le 30\sqrt{c} n \le 30\sqrt{n(k+1)}$. In the polynomial regime, it is proved that for any $\frac{1}{3} \le \alpha < 1$ and all sufficiently large $n$, if $k = \lfloor n^\alpha \rfloor$, then $f(k ,n) \le 9n^{(1+\alpha)/2} \le 9\sqrt{n(k+1)}$.

It is proved that for integers $b, r$ such that $3 \leq b < r \leq \binom{b+1}{2} - 1$, there exists a red/blue edge-colored graph such that the red degree of every vertex is $r$, the blue degree of every vertex is $b$, yet in the closed neighborhood of every vertex there are more blue edges than red edges. The upper bound $r \le \binom{b+1}{2}-1$ is best possible for any $b \ge 3$. We further extend this theorem to more than two colours, and to larger neighbourhoods. A useful result required in some of our proofs, of independent interest, is that for integers $r,t$ such that $0 \leq t \le \frac{r^2}{2} - 5r^{3/2}$, there exists an $r$-regular graph in which each open neighborhood induces precisely $t$ edges. Several explicit constructions are introduced and relationships with constant linked graphs, $(r,b)$-regular graphs and vertex transitive graphs are revealed.

An {\it inversion} of a tournament $T$ is obtained by reversing the direction of all edges with both endpoints in some set of vertices. Let ${\rm inv}_k(T)$ be the minimum length of a sequence of inversions using sets of size at most $k$ that result in the transitive tournament. Let ${\rm inv}_k(n)$ be the maximum of ${\rm inv}_k(T)$ taken over $n$-vertex tournaments. It is well-known that ${\rm inv}_2(n)=(1+o(1))n^2/4$ and it was recently proved by Alon et al. that ${\rm inv}(n):={\rm inv}_{n}(n)=n(1+o(1))$. In these two extreme cases ($k=2$ and $k=n$), random tournaments are asymptotically extremal objects. It is proved that the random tournament {\em does not} asymptotically attain ${\rm inv}_k(n)$ when $k \ge k_0$ and conjectured that ${\rm inv}_3(n)$ is (only) attained by (quasi) random tournaments. It is further proved that $(1+o(1)){\rm inv}_3(n)/n^2 \in [\frac{1}{12}, 0.0992)$ and $(1+o(1)){\rm inv}_k(n)/n^2 \in [\frac{1}{2k(k-1)}+\delta_k, \frac{1}{2 \lfloor k^2/2 \rfloor}-\epsilon_k]$ where $\epsilon_k > 0$ for all $k \ge 3$ and $\delta_k > 0$ for all $k \ge k_0$.

For every fixed $k \ge 4$, it is proved that if an $n$-vertex directed graph has at most $t$ pairwise arc-disjoint directed $k$-cycles, then there exists a set of at most $\frac{2}{3}kt+ o(n^2)$ arcs that meets all directed $k$-cycles and that the set of $k$-cycles admits a fractional cover of value at most $\frac{2}{3}kt$. It is also proved that the ratio $\frac{2}{3}k$ cannot be improved to a constant smaller than $\frac{k}{2}$. For $k=5$ the constant $2k/3$ is improved to $25/8$ and for $k=3$ it was recently shown by Cooper et al. that the constant can be taken to be $9/5$. The result implies a deterministic polynomial time $\frac{2}{3}k$-approximation algorithm for the directed $k$-cycle cover problem, improving upon a previous $(k{-}1)$-approximation algorithm of Kortsarz et al. More generally, for every directed graph $H$ we introduce a graph parameter $f(H)$ for which it is proved that if an $n$-vertex directed graph has at most $t$ pairwise arc-disjoint $H$-copies, then there exists a set of at most $f(H)t+ o(n^2)$ arcs that meets all $H$-copies and that the set of $H$-copies admits a fractional cover of value at most $f(H)t$. It is shown that for almost all $H$ it holds that $f(H) \approx |E(H)|/2$ and that for every $k$-vertex tournament $H$ it holds that $f(H) \le \lfloor k^2/4 \rfloor$., Comment: to appear in SIAM Journal on Discrete Mathematics

We present new algorithms for counting and detecting small tournaments in a given tournament. In particular, it is proved that every tournament on four vertices (there are four) can be detected in $O(n^2)$ time and counted in $O(n^\omega)$ time where $\omega < 2.373$ is the matrix multiplication exponent. It is also proved that any tournament on five vertices (there are $12$) can be counted in $O(n^{\omega+1})$ time. As for lower-bounds, we prove that for almost all $k$-vertex tournaments, the complexity of the detection problem is not easier than the complexity of the corresponding well-studied counting problem for {\em undirected cliques} of order $k-O(\log k)$.

In a recent breakthrough, Gilmer proved the union closed conjecture up to a constant factor. Using Gilmer's method and additional ideas, Chase and Lovett proved an optimal result for almost union-closed set systems. Here that result is extended to higher order unions.

For a graph $G$, let $c_k(G)$ be the number of spanning trees of $G$ with maximum degree at most $k$. For $k \ge 3$, it is proved that every connected $n$-vertex $r$-regular graph $G$ with $r \ge \frac{n}{k+1}$ satisfies $$ c_k(G)^{1/n} \ge (1-o_n(1)) r \cdot z_k $$ where $z_k > 0$ approaches $1$ extremely fast (e.g. $z_{10}=0.999971$). The minimum degree requirement is essentially tight as for every $k \ge 2$ there are connected $n$-vertex $r$-regular graphs $G$ with $r=\lfloor n/(k+1) \rfloor -2$ for which $c_k(G)=0$. Regularity may be relaxed, replacing $r$ with the geometric mean of the degree sequence and replacing $z_k$ with $z_k^* > 0$ that also approaches $1$, as long as the maximum degree is at most $n(1-(3+o_k(1))\sqrt{\ln k/k})$. The same holds with no restriction on the maximum degree as long as the minimum degree is at least $\frac{n}{k}(1+o_k(1))$., Comment: 25 pages, to appear in Random Structures & Algorithms

Let $H$ be a directed acyclic graph other than a rooted star. It is known that there are constants $c(H)$ and $C(H)$ such that the following holds for the complete directed graph $D_n$. There are at most $C\log n$ directed acyclic subgraphs of $D_n$ that cover every $H$-copy of $D_n$, while fewer than $c\log n$ directed acyclic subgraphs of $D_n$ do not cover all $H$-copies. Here this dichotomy is considerably strengthened. Let ${\vec G}(n,p)$ denote the random directed graph. The {\em fractional arboricity} of $H$ is $a(H) = max \{\frac{|E(H')|}{|V(H')|-1}\}$, where the maximum is over all non-singleton subgraphs of $H$. If $a(H) = \frac{|E(H)|}{|V(H)|-1}$ then $H$ is {\em totally balanced}. Complete graphs, complete multipartite graphs, cycles, trees, and, in fact, almost all graphs, are totally balanced. It is proved: 1) Let $H$ be a dag with $h$ vertices and $m$ edges other than a rooted star. For every $a^* > a(H)$ there exists $c^* = c^*(a^*,H) > 0$ such that almost surely $G \sim {\vec G}(n,n^{-1/a^*})$ has the property that every set $X$ of at most $c^*\log n$ directed acyclic subgraphs of $G$ does not cover all $H$-copies of $G$. Moreover, there exists $s(H) = m/2 + O(m^{4/5}h^{1/5})$ such that the following stronger assertion holds for any such $X$: There is an $H$-copy in $G$ that has no more than $s(H)$ of its edges covered by each element of $X$. 2) If $H$ is totally balanced then for every $0 < a^* < a(H)$, almost surely $G \sim {\vec G}(n,n^{-1/a^*})$ has a single directed acyclic subgraph that covers all its $H$-copies. As for the first result, note that if $h=o(m)$ then $s(H)=(1+o_m(1))m/2$ is about half of the edges of $H$. In fact, for infinitely many $H$ it holds that $s(H)=m/2$, optimally. As for the second result, the requirement that $H$ is totally balanced cannot, generally, be relaxed.

Let $H_r(n,p)$ denote the maximum number of Hamiltonian cycles in an $n$-vertex $r$-graph with density $p \in (0,1)$. The expected number of Hamiltonian cycles in the random $r$-graph model $G_r(n,p)$ is $E(n,p)=p^n(n-1)!/2$ and in the random graph model $G_r(n,m)$ with $m=p\binom{n}{r}$ it is, in fact, slightly smaller than $E(n,p)$. For graphs, $H_2(n,p)$ is proved to be only larger than $E(n,p)$ by a polynomial factor and it is an open problem whether a quasi-random graph with density $p$ can be larger than $E(n,p)$ by a polynomial factor. For hypergraphs (i.e. $r \ge 3$) the situation is drastically different. For all $r \ge 3$ it is proved that $H_r(n,p)$ is larger than $E(n,p)$ by an {\em exponential} factor and, moreover, there are quasi-random $r$-graphs with density $p$ whose number of Hamiltonian cycles is larger than $E(n,p)$ by an exponential factor.

Let $T$ be an arbitrary phylogenetic tree with $n$ leaves. It is well-known that the average quartet distance between two assignments of taxa to the leaves of $T$ is $\frac 23 \binom{n}{4}$. However, a longstanding conjecture of Bandelt and Dress asserts that $(\frac 23 +o(1))\binom{n}{4}$ is also the {\em maximum} quartet distance between two assignments. While Alon, Naves, and Sudakov have shown this indeed holds for caterpillar trees, the general case of the conjecture is still unresolved. A natural extension is when partial information is given: the two assignments are known to coincide on a given subset of taxa. The partial information setting is biologically relevant as the location of some taxa (species) in the phylogenetic tree may be known, and for other taxa it might not be known. What can we then say about the average and maximum quartet distance in this more general setting? Surprisingly, even determining the {\em average} quartet distance becomes a nontrivial task in the partial information setting and determining the maximum quartet distance is even more challenging, as these turn out to be dependent of the structure of $T$. In this paper we prove nontrivial asymptotic bounds that are sometimes tight for the average quartet distance in the partial information setting. We also show that the Bandelt and Dress conjecture does not generally hold under the partial information setting. Specifically, we prove that there are cases where the average and maximum quartet distance substantially differ.

The study of problems concerning subdivisions of graphs has a rich history in extremal combinatorics. Confirming a conjecture of Burr and Erd\H{o}s, Alon proved in 1994 that subdivided graphs have linear Ramsey numbers. Later, Alon, Krivelevich and Sudakov showed that every $n$-vertex graph with at least $\varepsilon n^2$ edges contains a $1$-subdivision of the complete graph on $c_{\varepsilon}\sqrt{n}$ vertices, resolving another old conjecture of Erd\H{o}s. In this paper we consider the directed analogue of these problems and show that every tournament on at least $(2+o(1))k^2$ vertices contains the 1-subdivision of a transitive tournament on $k$ vertices. This is optimal up to a multiplicative factor of 4 and confirms a conjecture of Gir\~ao, Popielarz and Snyder., Comment: 6 pages, improved constant, author list updated

The separation dimension of a hypergraph $G$ is the smallest natural number $d$ for which there is an embedding of $G$ into $\mathbb{R}^d$, such that any pair of disjoint edges is separated by some hyperplane normal to one of the axes. The perfect separation dimension further requires that any pair of disjoint edges is separated by the same amount of such (pairwise nonparallel) hyperplanes. While it is known that for any fixed $r \ge 2$, the separation dimension of any $n$-vertex $r$-graph is $O(\log n)$, the perfect separation dimension is much larger. In fact, no polynomial upper-bound for the perfect separation dimension of $r$-uniform hypergraphs is known. In our first result we essentially resolve the case $r=2$, i.e. graphs. We prove that the perfect separation dimension of $K_n$ is linear in $n$, up to a small polylogarithmic factor. In fact, we prove it is at least $n/2-1$ and at most $n(\log n)^{1+o(1)}$. Our second result proves that the perfect separation dimension of almost all graphs is also linear in $n$, up to a logarithmic factor. This follows as a special case of a more general result showing that the perfect separation dimension of the random graph $G(n,p)$ is w.h.p. $\Omega(n p /\log n)$ for a wide range of values of $p$, including all constant $p$. Finally, we prove that significantly relaxing perfection to just requiring that any pair of disjoint edges of $K_n$ is separated the same number of times up to a difference of $c \log n$ for some absolute constant $c$, still requires the dimension to be $\Omega(n)$. This is perhaps surprising as it is known that if we allow a difference of $7\log_2 n$, then the dimension reduces to $O(\log n)$., Comment: 22 pages, to appear in J. Combin. Des

We look at structures that must be removed (or reversed) in order to make acyclic a given oriented graph. For a directed acyclic graph $H$ and an oriented graph $G$, let $f_H(G)$ be the maximum number of pairwise disjoint copies of $H$ that can be found in {\em all} feedback arc sets of $G$. In particular, to make $G$ acyclic, one must remove (or reverse) $f_H(G)$ pairwise disjoint copies of $H$. Most intriguing is the case where $H$ is a $k$-clique, where the parameter is denoted by $f_k(G)$. Determining $f_k(G)$ for arbitrary $G$ seems challenging. Here we determine $f_k(G)$ precisely for almost all $k$-partite tournaments. Let $s(G)$ denote the size of the smallest vertex class of a $k$-partite tournament $G$. We prove that for all sufficiently large $s=s(G)$, a random $k$-partite tournament $G$ satisfies $f_k(G) = s(G)-k+1$ almost surely. In particular, as the title states, $f_k(G) = \lfloor n/k\rfloor-k+1$ almost surely, where $G$ is a random orientation of the Tur\'an graph $T(n,k)$., Comment: 23 pages

A $[k,n,1]$-graph is a $k$-partite graph with parts of order $n$ such that the bipartite graph induced by any pair of parts is a matching. An independent transversal in such a graph is an independent set that intersects each part in a single vertex. A factor of independent transversals is a set of $n$ pairwise-disjoint independent transversals. Let $f(k)$ be the smallest integer $n_0$ such that every $[k,n,1]$-graph has a factor of independent transversals assuming $n \ge n_0$. Several known conjectures imply that for $k \ge 2$, $f(k)=k$ if $k$ is even and $f(k)=k+1$ if $k$ is odd. While a simple greedy algorithm based on iterating Hall's Theorem shows that $f(k) \le 2k-2$, no better bound is known and in fact, there are instances showing that the bound $2k-2$ is tight for the greedy algorithm. Here we significantly improve upon the greedy algorithm bound and prove that $f(k) \le 1.78k$ for all $k$ sufficiently large, answering a question of MacKeigan., Comment: Final version with added reference

A vertex labeling of a hypergraph is sum distinguishing if it uses positive integers and the sums of labels taken over the distinct hyperedges are distinct. Let s(H) be the smallest integer N such that there is a sum-distinguishing labeling of H with each label at most N. The largest value of s(H) over all hypergraphs on n vertices and m hyperedges is denoted s(n,m). We prove that s(n,m) is almost-quadratic in m as long as m is not too large. More precisely, the following holds: If n < m < n^{O(1)} then s(n,m)= m^2/w(m), where w(m) is a function that goes to infinity and is smaller than any polynomial in m. The parameter s(n,m) has close connections to several other graph and hypergraph functions, such as the irregularity strength of hypergraphs. Our result has several applications, notably: 1. We answer a question of Gyarfas et al. whether there are n-vertex hypergraphs with irregularity strength greater than 2n. In fact we show that there are n-vertex hypergraphs with irregularity strength at least n^{2-o(1)}. 2. Our results imply that s*(n)=n^2/w(n) where s*(n) is the distinguishing closed-neighborhood number, i.e., the smallest integer N such that any n-vertex graph allows for a vertex labeling with positive integers at most N so that the sums of labels on distinct closed neighborhoods of vertices are distinct.

Since counting subgraphs in general graphs is, by and large, a computationally demanding problem, it is natural to try and design fast algorithms for restricted families of graphs. One such family that has been extensively studied is that of graphs of bounded degeneracy (e.g., planar graphs). This line of work, which started in the early 80's, culminated in a recent work of Gishboliner et al., which highlighted the importance of the task of counting homomorphic copies of cycles (i.e., cyclic walks) in graphs of bounded degeneracy. Our main result in this paper is a surprisingly tight relation between the above task and the well-studied problem of detecting (standard) copies of directed cycles in general directed graphs. More precisely, we prove the following: 1. One can compute the number of homomorphic copies of $C_{2k}$ and $C_{2k+1}$ in $n$-vertex graphs of bounded degeneracy in time $\tilde{O}(n^{d_{k}})$, where the fastest known algorithm for detecting directed copies of $C_k$ in general $m$-edge digraphs runs in time $\tilde{O}(m^{d_{k}})$. 2. Conversely, one can transform any $O(n^{b_{k}})$ algorithm for computing the number of homomorphic copies of $C_{2k}$ or of $C_{2k+1}$ in $n$-vertex graphs of bounded degeneracy, into an $\tilde{O}(m^{b_{k}})$ time algorithm for detecting directed copies of $C_k$ in general $m$-edge digraphs. We emphasize that our first result does not use a black-box reduction (as opposed to the second result which does). Instead, we design an algorithm for computing the number of $C_k$-homomorphisms in degenerate graphs and show that one part of its analysis can be reduced to the analysis of the fastest known algorithm for detecting directed cycles in general digraphs, which was carried out in a recent breakthrough of Dalirrooyfard, Vuong and Vassilevska Williams.

A shortcut of a directed path $v_1 v_2 \cdots v_n$ is an edge $v_iv_j$ with $j > i+1$. If $j = i+2$ the shortcut is called a hop. If all hops are present, the path is called hop complete, so the path and its hops form a square of a path. We prove that every tournament with $n \ge 4$ vertices has a Hamiltonian path with at least $(4n-10)/7$ hops, and has a hop complete path of order at least $n^{0.295}$. A spanning binary tree of a tournament is a spanning shortcut tree if for every vertex of the tree, all its left descendants are in-neighbors and all its right descendants are out-neighbors. It is well-known that every tournament contains a spanning shortcut tree. The number of shortcuts of a shortcut tree is the number of shortcuts of its unique induced Hamiltonian path. Let $t(n)$ denote the largest integer such that every tournament with $n$ vertices has a spanning shortcut tree with at least $t(n)$ shortcuts. We almost determine the asymptotic growth of $t(n)$ as it is proved that $\Theta(n\log^2n) \ge t(n)-\frac{1}{2}\binom{n}{2} \ge \Theta(n \log n)$.

Let ${\mathcal C}$ be a proper minor-closed family of graphs. We present a randomized algorithm that given a graph $G \in {\mathcal C}$ with $n$ vertices, finds a simple cycle of size $k$ in $G$ (if exists) in $2^{O(k)}n$ time. The algorithm applies to both directed and undirected graphs. In previous linear time algorithms for this problem, the runtime dependence on $k$ is super-exponential. The algorithm can be derandomized yielding a $2^{O(k)}n\log n$ time algorithm.

For a tournament $H$ with $h$ vertices, its typical density is $h!2^{-\binom{h}{2}}/aut(H)$, i.e. this is the expected density of $H$ in a random tournament. A family ${\mathcal F}$ of $h$-vertex tournaments is {\em dominant} if for all sufficiently large $n$, there exists an $n$-vertex tournament $G$ such that the density of each element of ${\mathcal F}$ in $G$ is larger than its typical density by a constant factor. Characterizing all dominant families is challenging already for small $h$. Here we characterize several large dominant families for every $h$. In particular, we prove the following for all $h$ sufficiently large: (i) For all tournaments $H^*$ with at least $5\log h$ vertices, the family of all $h$-vertex tournaments that contain $H^*$ as a subgraph is dominant. (ii) The family of all $h$-vertex tournaments whose minimum feedback arc set size is at most $\frac{1}{2}\binom{h}{2}-h^{3/2}\sqrt{\ln h}$ is dominant. For small $h$, we construct a dominant family of $6$ (i.e. $50\%$ of the) tournaments on $5$ vertices and dominant families of size larger than $40\%$ for $h=6,7,8,9$. For all $h$, we provide an explicit construction of a dominant family which is conjectured to obtain an absolute constant fraction of the tournaments on $h$ vertices. Some additional intriguing open problems are presented., Comment: To appear in Journal of Combinatorics

An $(n,k)$ sequence covering array is a set of permutations of $[n]$ such that each sequence of $k$ distinct elements of $[n]$ is a subsequence of at least one of the permutations. An $(n,k)$ sequence covering array is perfect if there is a positive integer $\lambda$ such that each sequence of $k$ distinct elements of $[n]$ is a subsequence of precisely $\lambda$ of the permutations. While relatively close upper and lower bounds for the minimum size of a sequence covering array are known, this is not the case for perfect sequence covering arrays. Here we present new nontrivial bounds for the latter. In particular, for $k=3$ we obtain a linear lower bound and an almost linear upper bound.

Let $F_k$ be the set of graphs on $k$ vertices. For a graph $G$, a $k$-decomposition is a set of induced subgraphs of $G$, each isomorphic to an element of $F_k$, such that each pair of vertices of $G$ is in exactly one element of the set. A fundamental result of Wilson is that for all $n=|V(G)|$ sufficiently large, $G$ has a $k$-decomposition if and only if $G$ is $k$-divisible. Let ${\bf v} \in {\mathbb R}^{|F_k|}$ be indexed by $F_k$. For a $k$-decomposition $L$ of $G$, let $\nu_{\bf v}(L) = \sum_{F \in F_k} {\bf v}_F d_{L,F}$ where $d_{L,F}$ is the fraction of elements of $L$ isomorphic to $F$. Let $\nu_{\bf v}(G) = \max_{L} \nu_{\bf v}(L)$ and $\nu_{\bf v}(n)=\min\{\nu_{\bf v}(G):|V(G)|=n\}$. It is not difficult to prove that the sequence $\nu_{\bf v}(n)$ has a limit so let $\nu_{\bf v} = \lim_{n \rightarrow \infty} \nu_{\bf v}(n)$. Replacing $k$-decompositions with their fractional relaxations, one obtains the (polynomial time computable) fractional analogue $\nu_{\bf v}^*(G)$ and corresponding fractional values $\nu^*_{\bf v}(n)$ and $\nu^*_{\bf v}$. Our first main result is that for each ${\bf v} \in {\mathbb R}^{|F_k|}$ $$ \nu_{\bf v} = \nu^*_{\bf v}\;. $$ Further, there is a polynomial time algorithm that produces a decomposition $L$ of a $k$-decomposable graph such that $\nu_{\bf v}(L) \ge \nu_{\bf v} - o_n(1)$. A similar result holds when $F_k$ is the family of all tournaments on $k$ vertices and when $F_k$ is the family of all edge-colorings of $K_k$. We use these results to obtain new and improved bounds on several decomposition results. For example, we prove that every $n$-vertex tournament which is $3$-divisible has a triangle decomposition in which the number of directed triangles is less than $0.0222n^2(1+o(1))$ and that every $5$-decomposable $n$-vertex graph has a $5$-decomposition in which the fraction of cycles of length $5$ is $o_n(1)$., Comment: 30 pages

Erd\H{o}s and Hanani proved that for every fixed integer $k \ge 2$, the complete graph $K_n$ can be almost completely packed with copies of $K_k$; that is, $K_n$ contains pairwise edge-disjoint copies of $K_k$ that cover all but an $o_n(1)$ fraction of its edges. Equivalently, elements of the set $\C(k)$ of all red-blue edge colorings of $K_k$ can be used to almost completely pack every red-blue edge coloring of $K_n$. The following strengthening of the aforementioned Erd\H{o}s-Hanani result is considered. Suppose $\C' \subset \C(k)$. Is it true that we can use elements only from $\C'$ and almost completely pack every red-blue edge coloring of $K_n$? An element $C \in \C(k)$ is {\em avoidable} if $\C'=\C(k) \setminus C$ has this property and a subset ${\cal F} \subset \C(k)$ is avoidable if $\C'=\C(k) \setminus {\cal F}$ has this property. It seems difficult to determine all avoidable graphs as well as all avoidable families. We prove some nontrivial sufficient conditions for avoidability. Our proofs imply, in particular, that (i) almost all elements of $\C(k)$ are avoidable (ii) all Eulerian elements of $\C(k)$ are avoidable and, in fact, the set of all Eulerian elements of $\C(k)$ is avoidable.

Erd\H{o}s, Fajtlowicz and Staton asked for the least integer $f(k)$ such that every graph with more than $f(k)$ vertices has an induced regular subgraph with at least $k$ vertices. Here we consider the following relaxed notions. Let $g(k)$ be the least integer such that every graph with more than $g(k)$ vertices has an induced subgraph with at least $k$ repeated degrees and let $h(k)$ be the least integer such that every graph with more than $h(k)$ vertices has an induced subgraph with at least $k$ maximum degree vertices. We obtain polynomial lower bounds for $h(k)$ and $g(k)$ and nontrivial linear upper bounds when the host graph has bounded maximum degree.

For an oriented graph $G$, let $f(G)$ denote the maximum chromatic number of an acyclic subgraph of $G$. Let $f(n)$ be the smallest integer such that every oriented graph $G$ with chromatic number larger than $f(n)$ has $f(G) > n$. Let $g(n)$ be the smallest integer such that every tournament $G$ with more than $g(n)$ vertices has $f(G) > n$. It is straightforward that $\Omega(n) \le g(n) \le f(n) \le n^2$. This paper provides the first nontrivial lower and upper bounds for $g(n)$. In particular, it is proved that $\frac{1}{4}n^{8/7} \le g(n) \le n^2-(2-\frac{1}{\sqrt{2}})n+2$. It is also shown that $f(2)=3$, i.e. every orientation of a $4$-chromatic graph has a $3$-chromatic acyclic subgraph. Finally, it is shown that a random tournament $G$ with $n$ vertices has $f(G) = \Theta(\frac{n}{\log n})$ whp.

Let $G$ and $H$ be graphs. We say that $P$ is an $H$-packing of $G$ if $P$ is a set of edge-disjoint copies of $H$ in $G$. An $H$-packing $P$ is maximal if there is no other $H$-packing of $G$ that properly contains $P$. Packings of maximum cardinality have been studied intensively, with several recent breakthrough results. Here, we consider minimum cardinality maximal packings. An $H$-packing $P$ is clumsy if it is maximal of minimum size. Let $cl(G,H)$ be the size of a clumsy $H$-packing of $G$. We provide nontrivial bounds for $cl(G,H)$, and in many cases asymptotically determine $cl(G,H)$ for some generic classes of graphs $G$ such as $K_n$ (the complete graph), $Q_n$ (the cube graph), as well as square, triangular, and hexagonal grids. We asymptotically determine $cl(K_n,H)$ for every fixed non-empty graph $H$. In particular, we prove that $$ cl(K_n, H) = \frac{\binom{n}{2}- ex(n,H)}{|E(H)|}+o(ex(n,H)),$$ where $ex(n,H)$ is the extremal number of $H$. A related natural parameter is $cov(G,H)$, that is the smallest number of copies of $H$ in $G$ (not necessarily edge-disjoint) whose removal from $G$ results in an $H$-free graph. While clearly $cov(G,H) \le cl(G,H)$, all of our lower bounds for $cl(G,H)$ apply to $cov(G,H)$ as well., Comment: 14 pages, 3 figures

Let $H$ be a graph on $h$ vertices. The number of induced copies of $H$ in a graph $G$ is denoted by $i_H(G)$. Let $i_H(n)$ denote the maximum of $i_H(G)$ taken over all graphs $G$ with $n$ vertices. Let $f(n,h) = \Pi_{i}^h a_i$ where $\sum_{i=1}^h a_i = n$ and the $a_i$ are as equal as possible. Let $g(n,h) = f(n,h) + \sum_{i=1}^h g(a_i,h)$. It is proved that for almost all graphs $H$ on $h$ vertices it holds that $i_H(n)=g(n,h)$ for all $n \le 2^{\sqrt{h}}$. More precisely, we define an explicit graph property ${\cal P}_h$ which, when satisfied by $H$, guarantees that $i_H(n)=g(n,h)$ for all $n \le 2^{\sqrt{h}}$. It is proved, in particular, that a random graph on $h$ vertices satisfies ${\cal P}_h$ with probability $1-o_h(1)$. Furthermore, all extremal $n$-vertex graphs yielding $i_H(n)$ in the aforementioned range are determined. We also prove a stability result. For $H \in {\cal P}_h$ and a graph $G$ with $n \le 2^{\sqrt{h}}$ vertices satisfying $i_H(G) \ge f(n,h)$, it must be that $G$ is obtained from a balanced blowup of $H$ by adding some edges inside the blowup parts. The {\em inducibility} of $H$ is $i_H = \lim_{n \rightarrow \infty} i_H(n)/\binom{n}{h}$. It is known that $i_H \ge h!/(h^h-h)$ for all graphs $H$ and that a random graph $H$ satisfies almost surely that $i_H \le h^{3\log h}h!/(h^h-h)$. We improve upon this upper bound almost matching the lower bound. It is shown that a graph $H$ which satisfies ${\cal P}_h$ has $i_H =(1+O(h^{-h^{1/3}}))h!/(h^h-h)$., Comment: 27 pages

Let ${\mathcal G}$ be an infinite family of connected graphs and let $k$ be a positive integer. We say that $k$ is ${\it forcing}$ for ${\mathcal G}$ if for all $G \in {\mathcal G}$ but finitely many, the following holds. Any $\{-1,1\}$-weighing of the edges of $G$ for which all connected subgraphs on $k$ edges are positively weighted implies that $G$ is positively weighted. Otherwise, we say that it is ${\it weakly~forcing}$ for ${\mathcal G}$ if any such weighing implies that the weight of $G$ is bounded from below by a constant. Otherwise we say that $k$ ${\it collapses}$ for ${\mathcal G}$. We classify $k$ for some of the most prominent classes of graphs, such as all connected graphs, all connected graphs with a given maximum degree and all connected graphs with a given average degree.

Suppose one needs to change the direction of at least $\epsilon n^2$ edges of an $n$-vertex tournament $T$, in order to make it $H$-free. A standard application of the regularity method shows that in this case $T$ contains at least $f^*_H(\epsilon)n^h$ copies of $H$, where $f^*_H$ is some tower-type function. It has long been observed that many graph/digraph problems become easier when assuming that the host graph is a tournament. It is thus natural to ask if the removal lemma becomes easier if we assume that the digraph $G$ is a tournament. Our main result here is a precise characterization of the tournaments $H$ for which $f^*_H(\epsilon)$ is polynomial in $\epsilon$, stating that such a bound is attainable if and only if $H$'s vertex set can be partitioned into two sets, each spanning an acyclic directed graph. The proof of this characterization relies, among other things, on a novel application of a regularity lemma for matrices due to Alon, Fischer and Newman, and on probabilistic variants of Ruzsa-Szemer\'edi graphs. We finally show that even when restricted to tournaments, deciding if $H$ satisfies the condition of our characterization is an NP-hard problem.

For an orientation $H$ with $n$ vertices, let $T(H)$ denote the maximum possible number of labeled copies of $H$ in an $n$-vertex tournament. It is easily seen that $T(H) \ge n!/2^{e(H)}$ as the latter is the expected number of such copies in a random tournament. For $n$ odd, let $R(H)$ denote the maximum possible number of labeled copies of $H$ in an $n$-vertex regular tournament. Adler et al. proved that, in fact, for $H=C_n$ the directed Hamilton cycle, $T(C_n) \ge (e-o(1))n!/2^{n}$ and it was observed by Alon that already $R(C_n) \ge (e-o(1))n!/2^{n}$. Similar results hold for the directed Hamilton path $P_n$. In other words, for the Hamilton path and cycle, the lower bound derived from the expectation argument can be improved by a constant factor. In this paper we significantly extend these results and prove that they hold for a larger family of orientations $H$ which includes all bounded degree Eulerian orientations and all bounded degree balanced orientations, as well as many others. One corollary of our method is that for any $k$-regular orientation $H$ with $n$ vertices, $T(H) \ge (e^k-o(1))n!/2^{e(H)}$ and in fact, for $n$ odd, $R(H) \ge (e^k-o(1))n!/2^{e(H)}$.

A path $v_1,v_2,\ldots,v_m$ in a graph $G$ is $degree$-$monotone$ if $deg(v_1) \leq deg(v_2) \leq \cdots \leq deg(v_m)$ where $deg(v_i)$ is the degree of $v_i$ in $G$. Longest degree-monotone paths have been studied in several recent papers. Here we consider the Ramsey type problem for degree monotone paths. Denote by $M_k(m)$ the minimum number $M$ such that for all $n \geq M$, in any $k$-edge coloring of $K_n$ there is some $1\leq j \leq k$ such that the graph formed by the edges colored $j$ has a degree-monotone path of order $m$. We prove several nontrivial upper and lower bounds for $M_k(m)$.

For a graph $H$, the {\em extremal number} $ex(n,H)$ is the maximum number of edges in a graph of order $n$ not containing a subgraph isomorphic to $H$. Let $\delta(H)>0$ and $\Delta(H)$ denote the minimum degree and maximum degree of $H$, respectively. We prove that for all $n$ sufficiently large, if $H$ is any graph of order $n$ with $\Delta(H) \le \sqrt{n}/200$, then $ex(n,H)={{n-1} \choose 2}+\delta(H)-1$. The condition on the maximum degree is tight up to a constant factor. This generalizes a classical result of Ore for the case $H=C_n$, and resolves, in a strong form, a conjecture of Glebov, Person, and Weps for the case of graphs. A counter-example to their more general conjecture concerning the extremal number of bounded degree spanning hypergraphs is also given.

One of the most basic results in graph theory states that every graph with at least two vertices has two vertices with the same degree. Since there are graphs without $3$ vertices of the same degree, it is natural to ask if for any fixed $k$, every graph $G$ is ``close'' to a graph $G'$ with $k$ vertices of the same degree. Our main result in this paper is that this is indeed the case. Specifically, we show that for any positive integer $k$, there is a constant $C=C(k)$, so that given any graph $G$, one can remove from $G$ at most $C$ vertices and thus obtain a new graph $G'$ that contains at least $\min\{k,|G|-C\}$ vertices of the same degree. Our main tool is a multidimensional zero-sum theorem for integer sequences, which we prove using an old geometric approach of Alon and Berman.

A minimum feedback arc set of a directed graph $G$ is a smallest set of arcs whose removal makes $G$ acyclic. Its cardinality is denoted by $\beta(G)$. We show that an Eulerian digraph with $n$ vertices and $m$ arcs has $\beta(G) \ge m^2/2n^2+m/2n$, and this bound is optimal for infinitely many $m, n$. Using this result we prove that an Eulerian digraph contains a cycle of length at most $6n^2/m$, and has an Eulerian subgraph with minimum degree at least $m^2/24n^3$. Both estimates are tight up to a constant factor. Finally, motivated by a conjecture of Bollob\'as and Scott, we also show how to find long cycles in Eulerian digraphs.

We present a $(1+\epsilon)$-approximation algorithm running in $O(f(\epsilon)\cdot n \log^4 n)$ time for finding the diameter of an undirected planar graph with non-negative edge lengths.

We present a new randomized algorithm for computing the diameter of a weighted directed graph. The algorithm runs in $\Ot(M^{\w/(\w+1)}n^{(\w^2+3)/(\w+1)})$ time, where $\w < 2.376$ is the exponent of fast matrix multiplication, $n$ is the number of vertices of the graph, and the edge weights are integers in $\{-M,...,0,...,M\}$. For bounded integer weights the running time is $O(n^{2.561})$ and if $\w=2+o(1)$ it is $\Ot(n^{7/3})$. This is the first algorithm that computes the diameter of an integer weighted directed graph polynomially faster than any known All-Pairs Shortest Paths (APSP) algorithm. For bounded integer weights, the fastest algorithm for APSP runs in $O(n^{2.575})$ time for the present value of $\w$ and runs in $\Ot(n^{2.5})$ time if $\w=2+o(1)$. For directed graphs with {\em positive} integer weights in $\{1,...,M\}$ we obtain a deterministic algorithm that computes the diameter in $\Ot(Mn^\w)$ time. This extends a simple $\Ot(n^\w)$ algorithm for computing the diameter of an {\em unweighted} directed graph to the positive integer weighted setting and is the first algorithm in this setting whose time complexity matches that of the fastest known Diameter algorithm for {\em undirected} graphs. The diameter algorithms are consequences of a more general result. We construct algorithms that for any given integer $d$, report all ordered pairs of vertices having distance {\em at most} $d$. The diameter can therefore be computed using binary search for the smallest $d$ for which all pairs are reported., Comment: revised to handle negative weights; faster algorithm for positive weights; added observation regarding the unweighted case

It is well known that a graph with $m$ edges can be made triangle-free by removing (slightly less than) $m/2$ edges. On the other hand, there are many classes of graphs which are hard to make triangle-free in the sense that it is necessary to remove roughly $m/2$ edges in order to eliminate all triangles. It is proved that dense graphs that are hard to make triangle-free, have a large packing of pairwise edge-disjoint triangles. In particular, they have more than $m(1/4+c\beta^2)$ pairwise edge-disjoint triangles where $\beta$ is the density of the graph and $c$ is an absolute constant. This improves upon a previous $m(1/4-o(1))$ bound which follows from the asymptotic validity of Tuza's conjecture for dense graphs. It is conjectured that such graphs have an asymptotically optimal triangle packing of size $m(1/3-o(1))$. The result is extended to larger cliques and odd cycles.

We prove that the sizes of the maximal dissociated subsets of a given finite subset of an abelian group differ by a logarithmic factor at most. On the other hand, we show that the set $\{0,1\}^n\seq\Z^n$ possesses a dissociated subset of size $\Ome(n\log n)$; since the standard basis of $\Z^n$ is a maximal dissociated subset of $\{0,1\}^n$ of size $n$, the result just mentioned is essentially sharp.

Let a_1,...,a_k satisfy a_1+...+a_k=1 and suppose a k-uniform hypergraph on n vertices satisfies the following property; in any partition of its vertices into k sets A_1,...,A_k of sizes a_1*n,...,a_k*n, the number of edges intersecting A_1,...,A_k is the number one would expect to find in a random k-uniform hypergraph. Can we then infer that H is quasi-random? We show that the answer is negative if and only if a_1=...=a_k=1/k. This resolves an open problem raised in 1991 by Chung and Graham [J. AMS '91]. While hypergraphs satisfying the property corresponding to a_1=...=a_k=1/k are not necessarily quasi-random, we manage to find a characterization of the hypergraphs satisfying this property. Somewhat surprisingly, it turns out that (essentially) there is a unique non quasi-random hypergraph satisfying this property. The proofs combine probabilistic and algebraic arguments with results from the theory of association schemes.

A {\em parametric weighted graph} is a graph whose edges are labeled with continuous real functions of a single common variable. For any instantiation of the variable, one obtains a standard edge-weighted graph. Parametric weighted graph problems are generalizations of weighted graph problems, and arise in various natural scenarios. Parametric weighted graph algorithms consist of two phases. A {\em preprocessing phase} whose input is a parametric weighted graph, and whose output is a data structure, the advice, that is later used by the {\em instantiation phase}, where a specific value for the variable is given. The instantiation phase outputs the solution to the (standard) weighted graph problem that arises from the instantiation. The goal is to have the running time of the instantiation phase supersede the running time of any algorithm that solves the weighted graph problem from scratch, by taking advantage of the advice. In this paper we construct several parametric algorithms for the shortest path problem. For the case of linear function weights we present an algorithm for the single source shortest path problem.

One of the main questions that arise when studying random and quasi-random structures is which properties P are such that any object that satisfies P "behaves" like a truly random one. In the context of graphs, Chung, Graham, and Wilson call a graph p-quasi-random} if it satisfies a long list of the properties that hold in G(n,p) with high probability, like edge distribution, spectral gap, cut size, and more. Our main result here is that the following holds for any fixed graph H: if the distribution of induced copies of H in a graph G is close (in a well defined way) to the distribution we would expect to have in G(n,p), then G is either p-quasi-random or p'-quasi-random, where p' is the unique non-trivial solution of a certain polynomial equation. We thus infer that having the correct distribution of induced copies of any single graph H is enough to guarantee that a graph has the properties of a random one. The proof techniques we develop here, which combine probabilistic, algebraic and combinatorial tools, may be of independent interest to the study of quasi-random structures.