Source author record

D. J. Grynkiewicz

D. J. Grynkiewicz appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

math.NT math.CA

Catalog footprint

What is connected

3works

2topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2012arXiv

The Large Davenport Constant I: Groups with a Cyclic, Index 2 Subgroup

Let $G$ be a finite group written multiplicatively. By a sequence over $G$, we mean a finite sequence of terms from $G$ which is unordered, repetition of terms allowed, and we say that it is a product-one sequence if its terms can be ordered so that their product is the identity element of $G$. The small Davenport constant $\mathsf d (G)$ is the maximal integer $\ell$ such that there is a sequence over $G$ of length $\ell$ which has no nontrivial, product-one subsequence. The large Davenport constant $\mathsf D (G)$ is the maximal length of a minimal product-one sequence---this is a product-one sequence which cannot be factored into two nontrivial, product-one subsequences. It is easily observed that $\mathsf d(G)+1\leq \mathsf D(G)$, and if $G$ is abelian, then equality holds. However, for non-abelian groups, these constants can differ significantly. Now suppose $G$ has a cyclic, index 2 subgroup. Then an old result of Olson and White (dating back to 1977) implies that $\mathsf d(G)=\frac12|G|$ if $G$ is non-cyclic, and $\mathsf d(G)=|G|-1$ if $G$ is cyclic. In this paper, we determine the large Davenport constant of such groups, showing that $\mathsf D(G)=\mathsf d(G)+|G'|$, where $G'=[G,G]\leq G$ is the commutator subgroup of $G$.

preprint2012arXiv

The Large Davenport Constant II: General Upper Bounds

Let $G$ be a finite group written multiplicatively. By a sequence over $G$, we mean a finite sequence of terms from $G$ which is unordered, repetition of terms allowed, and we say that it is a product-one sequence if its terms can be ordered so that their product is the identity element of $G$. The small Davenport constant $\mathsf d (G)$ is the maximal integer $\ell$ such that there is a sequence over $G$ of length $\ell$ which has no nontrivial, product-one subsequence. The large Davenport constant $\mathsf D (G)$ is the maximal length of a minimal product-one sequence---this is a product-one sequence which cannot be partitioned into two nontrivial, product-one subsequences. The goal of this paper is to present several upper bounds for $\mathsf D(G)$, including the following: $$\mathsf D(G)\leq {ll} \mathsf d(G)+2|G'|-1, & where $G'=[G,G]\leq G$ is the commutator subgroup; \frac34|G|, & if $G$ is neither cyclic nor dihedral of order $2n$ with $n$ odd; \frac{2}{p}|G|, & if $G$ is noncyclic, where $p$ is the smallest prime divisor of $|G|$; \frac{p^2+2p-2}{p^3}|G|, & if $G$ is a non-abelian $p$-group. As a main step in the proof of these bounds, we will also show that $\mathsf D(G)=2q$ when $G$ is a non-abelian group of order $|G|=pq$ with $p$ and $q$ distinct primes such that $p\mid q-1$.

preprint2010arXiv

Inverse Additive Problems for Minkowski Sumsets II

The Brunn-Minkowski Theorem asserts that $μ_d(A+B)^{1/d}\geq μ_d(A)^{1/d}+μ_d(B)^{1/d}$ for convex bodies $A,\,B\subseteq \R^d$, where $μ_d$ denotes the $d$-dimensional Lebesgue measure. It is well-known that equality holds if and only if $A$ and $B$ are homothetic, but few characterizations of equality in other related bounds are known. Let $H$ be a hyperplane. Bonnesen later strengthened this bound by showing $$μ_d(A+B)\geq (M^{1/(d-1)}+N^{1/(d-1)})^{d-1}(\frac{μ_d(A)}{M}+\frac{μ_d(B)}{N}),$$ where $M=\sup\{μ_{d-1}((\mathbf x+H)\cap A)\mid \mathbf x\in \R^d\}$ and $N=\sup\{μ_{d-1}((\mathbf y+H)\cap B)\mid \mathbf y\in \R^d\}$. Standard compression arguments show that the above bound also holds when $M=μ_{d-1}(π(A))$ and $N=μ_{d-1}(π(B))$, where $π$ denotes a projection of $\mathbb R^d$ onto $H$, which gives an alternative generalization of the Brunn-Minkowski bound. In this paper, we characterize the cases of equality in this later bound, showing that equality holds if and only if $A$ and $B$ are obtained from a pair of homothetic convex bodies by `stretching' along the direction of the projection, which is made formal in the paper. When $d=2$, we characterize the case of equality in the former bound as well.