MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

The upper Banach density of a subset $A\subset\mathbb N$ is defined $$d^*(A)=\limsup_{M-N\to\infty}\frac{|A\cap[M,M+1,\cdots,N]|}{M-N}$$ also the the upper density and lower density of a set $A$ is defined as following respectively $$\bar{d}(A)=\limsup_{n\to\infty}\frac{|A\cap[1,\cdots,n]|}{n}$$ $$\underline{d}(A)=\liminf_{n\to\infty}\frac{|A\cap[1,\cdots,n]|}{n}.$$ It is known that if $d^\ast(A)\gt 0$, then $A-A$ is $\Delta^*$. It means that $A-A$ has non-empty intersection with difference set of any sequence.

Now I want to know that Is there any subset $A\subset\mathbb N$ such that $\bar{d}(A-A)=1$?

share|cite|improve this question
Sorry, but this question is not appropriate to MO. What about $A = \mathbb N \setminus S$ with $S$ being any finite subset of $\mathbb N$? – Salvo Tringali Sep 29 '12 at 20:11
What is the assumption on $A$? As stated the question is trivial. Do you want to know if $\overline{d}(A-A)=1$ is possible even if $d^*(A)=0$? (the answer is yes, take $A$ as the set of natural numbers with coefficients $0$ and $1$ in their base $3$ expansion. Then $d^*(A)=0$ but $A-A=\mathbb{Z}$). There's also a typo in the definition of $\underline{d}(A)$. – Pablo Shmerkin Sep 29 '12 at 20:12
Thanks alot for your answer and I'm really sorry for my typos. But I want both $A$ and $A-A$ as a subset of natural numbers.I need a subset of $\mathbb N$ with zero Banach density such that $A-A=\{a_i-a_j: a_i>a_j\}\subset\mathbb N$ has full upper density. – maliheh Oct 10 '12 at 7:02
I think it is better that I ask in other way. I need a subset of natural numbers $P$ with $\bar{d}(P)=1$, but there is no $A\subset\mathbb N$ with positive Banach density such that $A-A\subset P$? In $\mathbb N$ or $\mathbb Z$ we can find other sequence by positive density where their difference set is in $\mathbb N$. – maliheh Oct 10 '12 at 7:07

Answering to the question formulated in the comments ("Is there a set $P$ with upper density $1$ such that $A-A$ is not contained in $P$ for any set $A$ with positive upper Banach density?"), note that for any $A$ with $\bar d(A)>0$, if $r>1/\bar d(A)$ then for any integers $a_1,...,a_r$ the total density of the sets $A-a_1,A-a_2,...,A-a_r$ is greater than $1$. Therefore $(A-a_i)\cap(A-a_j)$ is non-empty for some pair $(i,j)$. In other words, $A-A$ intersects the set of differences $\Delta(a_1,...,a_r)=\{a_i-a_j:1\leq i<j\leq r\}$.

Now construct $P$ as to avoid difference sets of arbitrarily large sets, but still have upper density $1$, for instance: $$P:=\mathbb{N}\setminus\left[\bigcup_{r=1}^\infty10^r\Delta(1,2,...,r)\right]$$ (here I used the usual notation, for $n\in\mathbb{N}$ and $X\subset\mathbb{N}$ denote by $nX:=\left\{nx:x\in X\right\}$). Thus $P$ has upper density $1$ (and actually Banach lower density $1$), but for any $A$ with positive Banach upper density $A-A\not\subset P$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.