1
$\begingroup$

Let $\|\,\,\,\|$ denote the Euclidean norm on $\mathbb R^2$. Let $\Lambda$ be a sublattice of $\mathbb Z^2$ and $m < M$ be positive real numbers. We say that a point $(x,y)$ in $\mathbb Z^2$ is primitive if $\gcd(x,y)=1$.

Define $\mathcal L_\Lambda(m,M)$ as a subcollection of primitive points $(x,y)$ in $\Lambda$ such that $m \leq \|(x,y)\|\leq M$, which is maximal with respect to the following property:

For any three distinct primitive points $(x_1,y_1),(x_2,y_2), (x_3,y_3) \in \mathcal L_\Lambda(m,M)$ there are no distinct $i,j,k\in \{1,2,3\}$ and no $\alpha,\beta \in \mathbb Z$ such that $(x_i,y_i) =\alpha (x_j,y_j)+\beta (x_k,y_k)$.

I am interested in obtaining "good" upper/lower bounds on the cardinality of $\mathcal L_\Lambda(m,M)$ in terms of $\Lambda, m, M$. This seems like a very natural question to ask, so I wanted to ask whether this problem was studied at all. Any suggestions on how to approach it would also be much appreciated.

$\endgroup$
1
  • $\begingroup$ Can you share the work you've already done concerning "possibly not so good" upper and lower bounds? Any progress in special cases (i.e., $\Lambda=\Bbb Z^2$ and $m=1$? $\Lambda=\Bbb Z^2$ and $m>M/2$?)? $\endgroup$ Oct 25, 2016 at 7:24

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.