On kissing numbers and spherical codes in high dimensions (Q1669035)

The celebrated kissing number problem asks for the maximal number \(K(d)\) of non-overlapping spheres of equal radius in \(d\)-dimensional Euclidean space touching the central sphere of the same radius. The origins of this problem go back to at least the famous ``13th ball'' discussion between Isaac Newton and David Gregory in 1694 about the 3-dimensional version of the problem, although the term ``kissing number'' comes from billiards and was only introduced in the second half of 20th century. The solution to the kissing number problem is only known in dimensions \(1, 2, 3, 4, 8, 24\) with all of these (except for the trivial cases \(1,2\)) being 20th century results. There are, however, known kissing number ranges in low dimensions, as well as general lower and upper bounds on \(K(d)\) that hold in all dimensions. The paper under review provides a nice brief overview of the literature and the main known results, while Conway \& Sloane's classical book can be consulted for further details [\textit{J. H. Conway} and \textit{N. J. A. Sloane}, Sphere packings, lattices and groups. With additional contributions by E. Bannai, R. E. Borcherds, J. Leech, S. P. Norton, A. M. Odlyzko, R. A. Parker, L. Queen and B. B. Venkov. 3rd ed. New York, NY: Springer (1999; Zbl 0915.52003)]. The main result of the paper under review is the new lower bound on \(K(d)\) of the form \[ K(d) \geq (1+o(1)) \frac{\sqrt{3\pi}}{8}\;\log \frac{3}{2\sqrt{2}}\;d^{3/2} \left( \frac{2}{\sqrt{3}} \right)^d. \] This constitutes a linear factor improvement to the best previously known bound which had \(\sqrt{d}\) instead of \(d^{3/2}\) with a slightly different constant. The contact points on the surface of the central sphere where neighboring spheres touch it form a kissing configuration, which is an example of a spherical code. In addition to the kissing number problem, the authors discuss more general spherical codes and produce a new lower bound on the maximal size of a spherical code in \(d\)-dimensional space, improving the previously known results by a linear factor in the dimension. The proofs use geometric and probabilistic techniques with the heart of the argument being two bounds on the expected size of a random spherical code.