9
$\begingroup$

Consider a dense sphere packing in $\mathbb{R}^n$, i.e. an arrangement of mutually disjoint solid open spheres, all of the same radius.

In dimensions $2, 3, 8,$ and $24$, it is known that lattice packings (packings where center of the spheres form a discrete subgroup of $\mathbb{R}^n$) are optimal.

It is widely believed that, in high enough dimensions, the best packings will be non-lattice packings. Is anything known about the best non-lattice packings? Do the methods of Viazovska et al. and Hales give values for the density of the best non-lattice packing? Is anything known (or conjectured) about the the ratio of the density of the best non-lattice packing to the density of the best lattice packing?

Improve this question
$\endgroup$
5
  • 6
    $\begingroup$ Just a comment which you're probably aware of: in 3 dimensions, there are non-lattice packings with the same density as lattice ones. That's part of the reason that it seems unlikely that Viazovska's approach will work in 3D. $\endgroup$
    – Ian Agol
    Sep 2, 2016 at 15:23
  • 1
    $\begingroup$ What qualifies for you as a non-lattice packing? If you take a lattice packing and remove one sphere, it has the same density and is not a lattice packing. Do you want it to be periodic? $\endgroup$
    – Yoav Kallus
    Sep 2, 2016 at 17:47
  • $\begingroup$ I guess I mean not a trivial deformation of a lattice packing. I understand that this a little vague, but I think that the question "what is the most dense packing in $\mathbb{R}^{24}$ that is constructed by doing something to the leech lattice" make sense. $\endgroup$
    – Matt Larson
    Sep 2, 2016 at 18:44
  • 2
    $\begingroup$ Do you perhaps mean "not constructed by doing something to the Leech lattice"? $\endgroup$
    – Noam D. Elkies
    Sep 3, 2016 at 4:32
  • $\begingroup$ @NoamD.Elkies Right, I cannot edit the comment anymore. $\endgroup$
    – Matt Larson
    Sep 3, 2016 at 15:08

1 Answer 1

Reset to default
4
$\begingroup$

I recently posted a preprint together with Alexei Andreanov in which we enumerate all the locally optimal 2-periodic sphere packings (also known as double-lattice packings) in dimensions up to $d=5$. This enumeration includes lattices, since they can be represented as 2-periodic arrangements by using a sublattice of index 2 as the unit cell. For the optimal packing density (among 2-periodic packings), we found that in $d=3$, there is one non-lattice that achieves it (the hexagonal close packing), in $d=4$ there are no non-lattices that achieve it, and for $d=5$ there are two. The highest locally optimal non-lattice double-lattice density in $d=4$ is the intriguing $\delta = 1/(2 \sqrt{9+4\sqrt{5}}) \approx 0.118$ (compared to $\delta = 1/8 = 0.125$ for the $D_4$ lattice).

Improve this answer
$\endgroup$

Your Answer

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

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