arXiv++ Combinatorics

Let $C_n$ be a cyclic group of order $n$. A sequence $S$ of length $\ell$ over $C_n$ is a sequence $S = a_1\boldsymbol\cdot a_2\boldsymbol\cdot \ldots\boldsymbol\cdot a_{\ell}$ of $\ell$ elements in $C_n$, where a repetition of elements is allowed and their order is disregarded. We say that $S$ is a zero-sum sequence if $Σ_{i=1}^{\ell} a_i = 0$ and that $S$ is a zero-sum free sequence if $S$ contains no zero-sum subsequence. Let $R$ be a random subset of $C_n$ obtained by choosing each element in $C_n$ independently with probability $p$. Let $N^R_{n-1-k}$ be the number of zero-sum free sequences of length $n-1-k$ in $R$. Also, let $N^R_{n-1-k,d}$ be the number of zero-sum free sequences of length $n-1-k$ having $d$ distinct elements in $R$. We obtain the expectation of $N^R_{n-1-k}$ and $N^R_{n-1-k,d}$ for $0\leq k\leq \big\lfloor \frac{n}{3} \big\rfloor$. We also show a concentration result on $N^R_{n-1-k}$ and $N^R_{n-1-k,d}$ when $k$ is fixed.

Two families, ${\mathcal A}$ and ${\mathcal B}$, of subsets of $[n]$ are cross $t$-intersecting if for every $A \in {\mathcal A}$ and $B \in {\mathcal B}$, $A$ and $B$ intersect in at least $t$ elements. For a real number $p$ and a family ${\mathcal A}$ the product measure $μ_p ({\mathcal A})$ is defined as the sum of $p^{|A|}(1-p)^{n-|A|}$ over all $A\in{\mathcal A}$. For every non-negative integer $r$, and for large enough $t$, we determine, for any $p$ satisfying $\frac r{t+2r-1}\leq p\leq\frac{r+1}{t+2r+1}$, the maximum possible value of $μ_p ({\mathcal A})μ_p ({\mathcal B})$ for cross $t$-intersecting families ${\mathcal A}$ and ${\mathcal B}$. In this paper we prove a stronger stability result which yields the above result.

We prove the following: Fix an integer $k\geq 2$, and let $T$ be a real number with $T\geq 1.5$. Let $\cH=(V,\cE_2\cup \cE_3\cup\dots\cup\cE_k)$ be a non-uniform hypergraph with the vertex set $V$ and the set $\cE_i$ of edges of size $i=2,\ldots , k$. Suppose that $\cH$ has no $2$-cycles (regardless of sizes of edges), and neither contains $3$-cycles nor $4$-cycles consisting of $2$-element edges. If the average degrees $t_i^{i-1} := i |\cE_i|/ |V|$ satisfy that $t_i^{i-1} \leq T^{i-1} (\ln T)^{\frac{k-i}{k-1}}$ for $i= 2, \dots , k$, then there exists a constant $C_k > 0$, depending only on $k$, such that $α(\cH)\geq C_k \frac{|V|}{T} (\ln T)^{\frac{1}{k-1}}$, where $α(\cH)$ denotes the independence number of $\cH$. This extends results of Ajtai, Komlós, Pintz, Spencer and Szemerédi and Duke, Rödl and the second author for uniform hypergraphs. As an application, we consider an anti-Ramsey type problem on non-uniform hypergraphs. Let $\cH=\cH(n;2,\ldots,\ell)$ be the hypergraph on the $n$-vertex set $V$ in which, for $s=2,\ldots,\ell$, each $s$-subset of $V$ is a hyperedge of $\cH$. Let $Δ$ be an edge-coloring of $\cH$ satisfying the following: (a) two hyperedges sharing a vertex have different colors; (b) two hyperedges with distinct size have different colors; (c) a color used for a hyperedge of size $s$ appears at most $u_s$ times. For such a coloring $Δ$, let $f_Δ(n;u_2,\ldots,u_{\ell})$ be the maximum size of a subset $U$ of $V$ such that each hyperedge of $\cH[U]$ has a distinct color, and let $f(n;u_2,\ldots,u_{\ell}):=\min_Δ f_Δ(n;u_2,\ldots,u_{\ell}).$ We determine $f(n;u_2,\ldots,u_{\ell})$ up to a multiplicative logarithm factor.

Ahlswede and Khachatrian's diametric theorem is a weighted version of their complete intersection theorem, itself an extension of the $t$-intersecting Erdős-Ko-Rado theorem. Their intersection theorem says that the maximum size of a family of subsets of $[n] = \{1, \dots, n\}$, every pair of which intersects in at least $t$ elements, is the size of certain trivially intersecting families proposed by Frankl. We address a cross intersecting version of their diametric theorem. Two families $\mathcal{A}$ and $\mathcal{B}$ of subsets of $[n]$ are {\em cross $t$-intersecting} if for every $A \in \mathcal{A}$ and $B \in \mathcal{B}$, $A$ and $B$ intersect in at least $t$ elements. The $p$-weight of a $k$ element subset $A$ of $[n]$ is $p^{k}(1-p)^{n-k}$, and the weight of a family $\mathcal{A}$ is the sum of the weights of its sets. The weight of a pair of families is the product of the weights of the families. The maximum $p$-weight of a $t$-intersecting family depends on the value of $p$. Ahlswede and Khachatrian showed that for $p$ in the range $[\frac{r}{t + 2r - 1}, \frac{r+1}{t + 2r + 1}]$, the maximum $p$-weight of a $t$-intersecting family is that of the family $\mathcal{F}^t_r$ consisting of all subsets of $[n]$ containing at least $t+r$ elements of the set $[t+2r]$. In a previous paper we showed a cross $t$-intersecting version of this for large $t$ in the case that $r = 0$. In this paper, we do the same in the case that $r = 1$. We show that for $p$ in the range $[\frac{1}{t + 1}, \frac{2}{t + 3}]$ the maximum $p$-weight of a cross $t$-intersecting pair of families, for $t \geq 200$, is achieved when both families are $\mathcal{F}^t_1$. Further, we show that except at the endpoints of this range, this is, up to isomorphism, the only pair of $t$-intersecting families achieving this weight.

When each vertex is assigned a set, the intersection graph generated by the sets is the graph in which two distinct vertices are joined by an edge if and only if their assigned sets have a nonempty intersection. An interval graph is an intersection graph generated by intervals in the real line. A chordal graph can be considered as an intersection graph generated by subtrees of a tree. In 1999, Karoński, Scheinerman and Singer-Cohen [Combin Probab Comput 8 (1999), 131--159] introduced a random intersection graph by taking randomly assigned sets. The random intersection graph $G(n,m;p)$ has $n$ vertices and sets assigned to the vertices are chosen to be i.i.d. random subsets of a fixed set $M$ of size $m$ where each element of $M$ belongs to each random subset with probability $p$, independently of all other elements in $M$. Fill, Scheinerman and Singer-Cohen [Random Struct Algorithms 16 (2000), 156--176] showed that the total variation distance between the random graph $G(n,m;p)$ and the Erdös-Rényi graph $G(n,\hat{p})$ tends to $0$ for any $0 \leq p=p(n) \leq 1$ if $m=n^α$, $α>6$, where $\hat{p}$ is chosen so that the expected numbers of edges in the two graphs are the same. In this paper, it is proved that the total variation distance still tends to $0$ for any $0 \leq p=p(n) \leq 1$ whenever $m \gg n^4$.

For positive integers $d$ and $n$, let $[n]^d$ be the set of all vectors $(a_1,a_2,\dots, a_d)$, where $a_i$ is an integer with $0\leq a_i\leq n-1$. A subset $S$ of $[n]^d$ is called a \emph{Sidon set} if all sums of two (not necessarily distinct) vectors in $S$ are distinct. In this paper, we estimate two numbers related to the maximum size of Sidon sets in $[n]^d$. First, let $\mathcal{Z}_{n,d}$ be the number of all Sidon sets in $[n]^d$. We show that $\log (\mathcal{Z}_{n,d})=Θ(n^{d/2})$, where the constants of $Θ$ depend only on $d$. Next, we estimate the maximum size of Sidon sets contained in a random set $[n]^d_p$, where $[n]^d_p$ denotes a random set obtained from $[n]^d$ by choosing each element independently with probability $p$.

For a family $\mathcal{F}$ of graphs, a graph $G$ is called \emph{$\mathcal{F}$-universal} if $G$ contains every graph in $\mathcal{F}$ as a subgraph. Let $\mathcal{F}_n(d)$ be the family of all graphs on $n$ vertices with maximum degree at most $d$. Dellamonica, Kohayakawa, Rödl and Ruciński showed that, for $d\geq 3$, the random graph $G(n,p)$ is $\mathcal{F}_n(d)$-universal with high probability provided $p\geq C\big(\frac{\log n}{n}\big)^{1/d}$ for a sufficiently large constant $C=C(d)$. In this paper we prove the missing part of the result, that is, the random graph $G(n,p)$ is $\mathcal{F}_n(2)$-universal with high probability provided $p\geq C\big(\frac{\log n}{n}\big)^{1/2}$ for a sufficiently large constant $C$.

Two families $\mathcal{A}$ and $\mathcal{B}$, of $k$-subsets of an $n$-set, are {\em cross $t$-intersecting} if for every choice of subsets $A \in \mathcal{A}$ and $B \in \mathcal{B}$ we have $|A \cap B| \geq t$. We address the following conjectured cross $t$-intersecting version of the Erd\H os--Ko--Rado Theorem: For all $n \geq (t+1)(k-t+1)$ the maximum value of $|\mathcal{A}||\mathcal{B}|$ for two cross $t$-intersecting families $\mathcal{A}, \mathcal{B} \subset\binom{[n]}{k}$ is $\binom{n-t}{k-t}^2$. We verify this for all $t \geq 14$ except finitely many $n$ and $k$ for each fixed $t$. Further, we prove uniqueness and stability results in these cases, showing, for instance, that the families reaching this bound are unique up to isomorphism. We also consider a {\em $p$-weight} version of the problem, which comes from the product measure on the power set of an $n$-set.

For a rational number $r>1$, a set $A$ of positive integers is called an $r$-multiple-free set if $A$ does not contain any solution of the equation $rx = y$. The extremal problem on estimating the maximum possible size of $r$-multiple-free sets contained in $[n]:={1,2,...,n}$ has been studied for its own interest in combinatorial number theory and application to coding theory. Let $a$, $b$ be positive integers such that $a<b$ and the greatest common divisor of $a$ and $b$ is 1. Wakeham and Wood showed that the maximum size of $(b/a)$-multiple-free sets contained in $[n]$ is $\frac{b}{b+1}n+O(\log n)$. In this paper we generalize this result as follows. For a real number $p\in (0,1)$, let $[n]_p$ be a set of integers obtained by choosing each element $i\in [n]$ randomly and independently with probability $p$. We show that the maximum possible size of $(b/a)$-multiple-free sets contained in $[n]_p$ is $\frac{b}{b+p}pn+O(\sqrt{pn}\log n \log \log n)$ with probability that goes to 1 as $n\to \infty$.

We prove that every simple connected graph with no $K_5$ minor admits a proper 4-coloring such that the neighborhood of each vertex $v$ having more than one neighbor is not monochromatic, unless the graph is isomorphic to the cycle of length 5. This generalizes the result by S.-J. Kim, S. J. Lee, and W.-J. Park on planar graphs.