arXiv++ Combinatorics

We derive an asymptotic formula for the number of solutions in a given subfield to certain system of equations over finite fields. As an application, we construct new families of maximal cliques in generalized Paley graphs. Given integers $d\ge2$ and $q \equiv 1 \pmod d$, we show that for each positive integer $m$ such that $\operatorname{rad}(m) \mid \operatorname{rad}(d)$, there are maximal cliques of size approximately $q/m$ in the $d$-Paley graph defined on $\mathbb{F}_{q^d}$. We also confirm a conjecture of Goryainov, Shalaginov, and the second author on the maximality of certain cliques in generalized Paley graphs, as well as an analogous conjecture of Goryainov for Peisert graphs.

Let $p$ be a prime and $n$ a positive integer such that $\sqrt{\frac p2} + 1 \leq n \leq \sqrt{p}$. For any arithmetic progression $A$ of length $n$ in $\mathbb{F}_p$, we establish an asymptotic formula for the number of directions determined by $A \times A \subset \mathbb{F}_p^2$. The key idea is to reduce the problem to counting the number of solutions to the bilinear Diophantine equation $ad+bc=p$ in variables $1\le a,b,c,d\le n$; our asymptotic formula for the number of solutions is of independent interest.

We adapt a number-theoretic technique of Yu to prove a purely analytic theorem: if f(x) is in L^1 and L^2, is nonnegative, and is supported on an interval of length I, then the supremum of the convolution f*f is at least 0.631 \| f \|_1^2 / I. This improves the previous bound of 0.591389 \| f \|_1^2 / I. Consequently, we improve the known bounds on several related number-theoretic problems. For a subset A of {1,2, ..., n}, let g be the maximum multiplicity of any element of the multiset {a+b: a,b in A}. Our main corollary is the inequality gn>0.631|A|^2, which holds uniformly for all g, n, and A.

Since addition is commutative but subtraction is not, the sumset S+S of a finite set S is predisposed to be smaller than the difference set S-S. In this paper, however, we show that each of the three possibilities (|S+S|>|S-S|, |S+S|=|S-S|, |S+S|<|S-S|) occur for a positive proportion of the subsets of {0, 1, ..., n-1}. We also show that the difference |S+S| - |S-S| can take any integer value, and we show that the expected number of omitted differences is asymptotically 6 while the expected number of missing sums is asymptotically 10. Other data and conjectures on the distribution of these quantities are also given.

A symmetric subset of the reals is one that remains invariant under some reflection z --> c-z. We consider, for any 0 < x <= 1, the largest real number D(x) such that every subset of $[0,1]$ with measure greater than x contains a symmetric subset with measure D(x). In this paper we establish upper and lower bounds for D(x) of the same order of magnitude: for example, we prove that D(x) = 2x - 1 for 11/16 <= x <= 1 and that 0.59 x^2 < D(x) < 0.8 x^2 for 0 < x <= 11/16. This continuous problem is intimately connected with a corresponding discrete problem. A set S of integers is called a B*[g] set if for any given m there are at most g ordered pairs (s_1,s_2) \in S \times S with s_1+s_2 = m; in the case g=2, these are better known as Sidon sets. Our lower bound on D(x) implies that every B*[g] set contained in \{1,2,...,n\} has cardinality less than 1.30036 \sqrt{gn}. This improves a result of Green for g >= 30. Conversely, we use a probabilistic construction of B*[g] sets to establish an upper bound on D(x) for small x.

We give explicit constructions of sets S with the property that for each integer k, there are at most g solutions to k=s_1+s_2, s_i\in S; such sets are called Sidon sets if g=2 and generalized Sidon sets if g\ge 3. We extend to generalized Sidon sets the Sidon-set constructions of Singer, Bose, and Ruzsa. We also further optimize Koulantzakis' idea of interleaving several copies of a Sidon set, extending the improvements of Cilleruelo & Ruzsa & Trujillo, Jia, and Habsieger & Plagne. The resulting constructions yield the largest known generalized Sidon sets in virtually all cases.

A symmetric subset of the reals is one that remains invariant under some reflection x --> c-x. Given 0 < x < 1, there exists a real number D(x) with the following property: if 0 < d < D(x), then every subset of [0,1] with measure x contains a symmetric subset with measure d, while if d > D(x), then there exists a subset of [0,1] with measure x that does not contain a symmetric subset with measure d. In this paper we establish upper and lower bounds for D(x) of the same order of magnitude: for example, we prove that D(x) = 2x - 1 for 11/16 < x < 1 and that 0.59 x^2 < D(x) < 0.8 x^2 for 0 < x < 11/16. This continuous problem is intimately connected with a corresponding discrete problem. A set S of integers is called a B*[g] set if for any given m there are at most g ordered pairs (s_1,s_2) \in S \times S with s_1+s_2=m; in the case g=2, these are better known as Sidon sets. We also establish upper and lower bounds of the same order of magnitude for the maximal possible size of a B*[g] set contained in {1,...,n}, which we denote by R(g,n). For example, we prove that R(g,n) < 1.31 \sqrt{gn} for all n > g > 1, while R(g,n) > 0.79 \sqrt{gn} for sufficiently large integers g and n. These two problems are so interconnected that both continuous and discrete tools can be applied to each problem with surprising effectiveness. The harmonic analysis methods and inequalities among various L^p norms we use to derive lower bounds for D(x) also provide uniform upper bounds for R(g,n), while the techniques from combinatorial and probabilistic number theory that we employ to obtain constructions of large B*[g] sets yield strong upper bounds for D(x).

In this paper we correct an analysis of the two-player perfect-information game Dukego given in Berlekamp, Conway, and Guy's Winning Ways for your Mathematical Plays (Chapter 19). In particular, we characterize the board dimensions that are fair, i.e., those for which the first player to move has a winning strategy.

Moser asked whether the collection of rectangles of dimensions 1 x 1/2, 1/2 x 1/3, 1/3 x 1/4, ..., whose total area equals 1, can be packed into the unit square without overlap, and whether the collection of squares of side lengths 1/2, 1/3, 1/4, ... can be packed without overlap into a rectangle of area pi^2/6-1. Computational investigations have been made into packing these collections into squares of side length 1+epsilon and rectangles of area pi^2/6-1+epsilon, respectively, and one can consider the apparently weaker question whether such packings are possible for every positive number epsilon. In this paper we establish a general theorem on sequences of geometrical packings that implies in particular that the ``for every epsilon'' versions of these two problems are actually equivalent to the original tiling problems.