Progress! (Apollonian Sphere Packings)

Since giving the recent talk, I had a bit of a breakthrough.  Two of the problems I mentioned are no longer open!  I now know (and can prove) the following:

Theorem: If a,b,c are nonzero integers, at most one of which is negative, then the numberN_3(a,b,c) of inequivalent occurrences of three mutually tangent spheres of curvature a,b,c is the same as the number of algebraic integers in {\mathbb{Q}}(\sqrt{-3}) with norm (ab+ac+bc), up to multiplication by a unit.  The bijection between algebraic integers with that norm and triples of spheres is natural and explicit.

There is a nice way to compute this number.  Define an arithmetic function \eta on the positive integers as follows.  If p=2 or p\equiv -1\pmod 6, then \eta(p^k)=1 if k is even and \eta(p^k)=0 if k is odd; if p\equiv 1 \pmod 6, then \eta(p^k)=k+1; \eta(3^k)=1 for all k.  Extend \eta to all positive integers by multiplicativity.  Then N_3(a,b,c)=\eta(ab+ac+bc).

Theorem: If a,b are nonzero integers, at most one of which is negative, then the numberN_3(a,b) of inequivalent occurrences of tangent spheres of curvature a,b is given by the formula N_3(a,b) = \left\lceil \frac{1}{12}\left(M_\star+3M_1+3M_3\right)\right\rceil + X.  Here M_\star is the number of solutions (u,v) to the congruence u^2+uv+v^2\equiv ab \pmod{a+b}, M_k is the number of solutions u to the congruence ku^2\equiv ab \pmod{a+b}, and X is 1 if 12|(a+b) and (a+b)|ab, otherwise 0.

(This strange formula comes from an application of Burnside’s Counting Lemma; as always in these problems, the trickiest part is keeping track of which solutions correspond to the same packing.)

The next natural problem in the progression, counting the total number of inequivalent occurrences of a given curvature in integer sphere packings, remains resistant to my current approach.

In the next week, I should have the corresponding result for counting occurrences of a given n-tuple of n-dimensional hyperspheres.  If I’m lucky, my 2- and 3-dimensional techniques will generalize to let me count occurrences of (n-1)-tuples in n dimensions, but I can’t be sure of that part yet.   Beyond that, I currently have no idea how to proceed.


Tags: , ,

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: