0
$\begingroup$

In the case of a square lattice, the exact number of points within a circle of radius r centered in the center is (see: http://mathworld.wolfram.com/GausssCircleProblem.html: $$N(r)=1+4Floor(r)+4 \sum_{x=1}^{Floor(r)}{Floor(\sqrt{r^2−x^2)}}$$ And in the case of an hexagonal lattice, I found in this post An exact counting solution for the number of points within a circle of radius $r$ centered on a lattice point in a $A_2$ hexagonal lattice that the number of points within a circle of radius r centered in the center is: $$ N(r)= \sum_{x = -Floor(\frac{r}{\sqrt{3}})}^{Floor(\frac{r}{\sqrt{3}})}( 1 + 2 Floor(\sqrt{r^2 - 3 x^2}) + \sum_{x = -Floor(\frac{r}{\sqrt{3}} + \frac{1}{2}) + \frac{1}{2}}^{Floor(\frac{r}{\sqrt{3}} + \frac{1}{2}) - \frac{1}{2}}( 2 Floor(\sqrt{r^2 - 3 x^2} + \frac{1}{2}). $$ And I checked this expression with the values in http://oeis.org/A053416 and I don't obtain the same values for an r. Can you guide me where I am wrong? In my research I want to obtain the number of the lattice points for square and hexagonal lattices in function of lattice constant and region size. This is the structure that I obtain numerrically I am new to this subject and I appreciate all the suggestions.

$\endgroup$
9
  • $\begingroup$ How can we guide you if you don't show your work? And anyway, this question might be better fit for Mathematics. $\endgroup$
    – Alex M.
    Commented Feb 18, 2022 at 16:38
  • $\begingroup$ From the mathematic point of view, the math behind this exact solutions for someone that doesn't not work in the field field (books). I did my structures numerically and now I need to find the number of points within a circle of radius r. For the square lattice the expression is checked, but what I find for the hexagonal lattice it's not correct or maybe I am wrong. I found this expression on a post here. I don't have the skills to check this expression so I would appreciate some help in this direction. $\endgroup$
    – Mihaela
    Commented Feb 18, 2022 at 16:49
  • $\begingroup$ Is it possible that the mathoverflow link and the oeis link aren't counting exactly the same thing? Maybe they are situating the lattice differently with respect to the origin, or maybe they are not using hexagons of the same size? $\endgroup$ Commented Feb 18, 2022 at 22:52
  • $\begingroup$ As is in showed in that post, https://mathoverflow.net/questions/110186/an-exact-counting-solution-for-the-number-of-points-within-a-circle-of-radius-r, it should give the same results. The only difference that I should take in the account is that in oeis link is expressed the number of points in function of the diameter of the circle and in the formula below it's expressed in function of the radius. But I took into account that. $\endgroup$
    – Mihaela
    Commented Feb 19, 2022 at 7:07
  • $\begingroup$ The formula above I think it not correct because in the case of radius 1, the number of points should be 7 and this is the number in oeis. But the person who has written this is claiming that it gives the results in oeis. $\endgroup$
    – Mihaela
    Commented Feb 19, 2022 at 7:07

1 Answer 1

5
$\begingroup$

By identifying the lattice points with numbers of the form $x - y\omega$, $\omega = e^{2\pi i / 3}$, $x, y \in \mathbb{Z}$, we find that we want to count Diophantine solutions to $x^2 + xy + y^2 \le r^2$. From $$\sum_{m, n \in \mathbb{Z}} q^{m^2 + mn + n^2} = 1 + 6 \sum_{n \ge 0} \left(\frac{q^{3n+1}}{1 - q^{3n+1}} - \frac{q^{3n+2}}{1-q^{3n+2}}\right)$$ (see e.g. Michael D. Hirschhorn (1999) Three Classical Results on Representations of a Number, Séminaire Lotharingien de Combinatoire, B42f) we get $$N(r) = 1 + 6 \sum_{k=0}^\infty \left\lfloor \frac{r^2}{3k+1} \right\rfloor - \left\lfloor \frac{r^2}{3k+2} \right\rfloor$$

$\endgroup$
11
  • $\begingroup$ Can you give me more references? This result is related to Eisenstein integer? How can we determine the multiple 6? And the "3k+1" and "3k+1"? I looked at Diophantine solutions and the material given by you, and some books. I need more basic understanding of this problems $\endgroup$
    – Mihaela
    Commented Mar 10, 2022 at 17:15
  • $\begingroup$ @Mihaela, yes, the lattice points are the Eisenstein integers. Normally we'd write $x + y\omega$, but the bijection $y \leftrightarrow -y$ is straightforward and puts the norm in the right form to directly apply the identity. The multiple 6 is unsurprising: the hexagonal lattice has order 6 rotation. The extraction of $N(r)$ from the generating function is by treating it as a difference of geometric series: $$\frac{x}{1-x} = x + x^2 + x^3 + \cdots$$ so that $k$ in the second identity takes the part of $n$ on the right-hand side of the first. $\endgroup$ Commented Mar 10, 2022 at 17:36
  • $\begingroup$ @Mihaela, I've put an old exposition of the Hirschhorn paper up on my website. $\endgroup$ Commented Mar 11, 2022 at 15:46
  • $\begingroup$ Can you get rid of infinity, on the sum, like in square lattice? $\endgroup$
    – Mihaela
    Commented Apr 15, 2022 at 19:19
  • $\begingroup$ In the square lattice $$N(r) = 1 + 4 \sum_{k=0}^\infty \left\lfloor \frac{r^2}{4k+1} \right\rfloor - \left\lfloor \frac{r^2}{4k+1} \right\rfloor$$ $\endgroup$
    – Mihaela
    Commented Apr 15, 2022 at 19:25

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .