Improved Delsarte bounds for spherical codes in small dimensions (Q2642026)

A spherical \((n,N,\alpha)\) code is a set of \(N\) unit vectors in \(n\)-dimensional Euclidean space such that the pairwise angular distance between the vectors is at least \(\alpha\). The central questions are maximizing either \(N\) or \(\alpha\) while fixing the other two parameters. The well known technique used by Delsarte which established bounds for binary codes and association schemes was extended by Delsarte, Goethals and Seidel to the case of spherical codes. The author extends this linear programming method for spherical codes and examines the particular case corresponding to the kissing number problem. Specifically upper bounds are improved for kissing numbers. Recent work by Musin in dimensions three and four is cast in this setting.