0
$\begingroup$

For high enough dimension $n$ there are lattices $L_n$ in $\mathbb{R}^n$ whose Voronoi partition's base regions encompass all but a negligible proportion of a $(1-\varepsilon)$-ball, and also nearly contain a $(1+\varepsilon)$-ball to all but a negligible proportion.[1] How much can such base regions behave like a ball in other ways?


One can define a "ball mod" as $mod_\ast(v)=(v/\|v\|)\cdot [\|v\| - \ \operatorname{round}(\|v\|/2)]$, which just reduces $v$ by lengths of 2 until it is in the unit ball at the origin.

One can also define "round" and "mod" operations over space up to a lattice $L$ and its Voronoi partition as: $R_L(v)=\arg\min_{\ell\in L}\|v-\ell\|$ and $mod_L(v)=v-R_L(v).$

Fixing radius $R>0$, for high enough dimension $n$, then does a lattice $L$ of the sort described above have that: $\|mod_\ast(v)-\operatorname{mod}_L(v)\|<\varepsilon$ for most $v\in B(0,R)$ ?


[1]: https://ieeexplore.ieee.org/abstract/document/1512416

$\endgroup$
  • $\begingroup$ I guess no, since I suspect there are not enough Voronoi regions adjacent to the base region for a lattice-mod-reduction of a vector just slightly outside of the base region to commonly approximately correspond to that vector's length-2 reduction. $\endgroup$ – enthdegree Nov 29 '18 at 23:58

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.