5
$\begingroup$

I am wondering if it is possible to produce an upper and/or lower bound on the number of integer lattice points that lie inside an $n$-dimensional ellipse.

That is, given an $n$-dimensional ellipsoid with a center at $c \in \mathbb{R}^n$: $$E(A,c,r) := \Big\{x \in \mathbb{R}^n ~\Big| ~(x-c)^T A (x-c) \leq r\Big\},$$

does there exist an upper and/or lower bound for the quantity:

$$\#(E(A,c,r)) := \big|E(A,c,r) \cap \mathbb{Z}^n \big|$$

If so, could someone cite the result / direct me towards the appropriate reference?

I'd appreciate any guidance. I'm really new to this area (and very unfamiliar with the terms). The only relevant result that I know is Minkowski's Convex Body theorem (though this just guarantees 3 lattice points as far as I know).

$\endgroup$
4
  • $\begingroup$ What kind of bound do you expect? $\endgroup$ – Fan Zheng Apr 23 '15 at 20:53
  • $\begingroup$ @FanZheng This might be naive since I'm still really new to all of this. That said, I am looking for a non-asymptotic upper and/or lower bound that is guaranteed to hold for any fixed $n$ (i.e. a lower bound that says you will have at least 10 lattice points in $E$ and at most 1000 lattice points in $E$). The bounds do not have to be related, and they can make use any of the quantities above (i.e. the volume of $E$). $\endgroup$ – Berk U. Apr 23 '15 at 21:12
  • $\begingroup$ U Do you mind if the lower bound is negative if $r$ is positive but small? If you try to turn an asymptotic bound to an absolute one, this is likely to happen. $\endgroup$ – Fan Zheng Apr 23 '15 at 21:19
  • $\begingroup$ @FanZheng No not at all. $\endgroup$ – Berk U. Apr 24 '15 at 18:37
6
$\begingroup$

Well, it depends on what you want to know. For homothetic images $t E$ of a fixed ellipsoid $E,$ the quantity is asymptotic to $t^n vol(E),$ where the error term is $O(t^{n-1}),$ with better bounds provable. But maybe you want to know something else?!

EDIT For non-asymptotic bounds, see this paper of Bentkus and Gotze (Acta Arith. 1997), which has copious references also.

$\endgroup$
2
  • $\begingroup$ Thanks for this! I added a comment above to clarify the type of bound that I was expecting... I'm not sure if it helps since I'm still new to all of this, but let me know and I will try to be more specific? $\endgroup$ – Berk U. Apr 23 '15 at 21:15
  • $\begingroup$ @BerkU. See the edit. $\endgroup$ – Igor Rivin Apr 23 '15 at 23:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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