English

Optimality of spherical codes via exact semidefinite programming bounds

Metric Geometry 2024-03-26 v1 Information Theory math.IT Optimization and Control

Abstract

We show that the spectral embeddings of all known triangle-free strongly regular graphs are optimal spherical codes (the new cases are 5656 points in 2020 dimensions, 5050 points in 2121 dimensions, and 7777 points in 2121 dimensions), as are certain mutually unbiased basis arrangements constructed using Kerdock codes in up to 10241024 dimensions (namely, 24k+22k+12^{4k} + 2^{2k+1} points in 22k2^{2k} dimensions for 2k52 \le k \le 5). As a consequence of the latter, we obtain optimality of the Kerdock binary codes of block length 6464, 256256, and 10241024, as well as uniqueness for block length 6464. We also prove universal optimality for 288288 points on a sphere in 1616 dimensions. To prove these results, we use three-point semidefinite programming bounds, for which only a few sharp cases were known previously. To obtain rigorous results, we develop improved techniques for rounding approximate solutions of semidefinite programs to produce exact optimal solutions.

Keywords

Cite

@article{arxiv.2403.16874,
  title  = {Optimality of spherical codes via exact semidefinite programming bounds},
  author = {Henry Cohn and David de Laat and Nando Leijenhorst},
  journal= {arXiv preprint arXiv:2403.16874},
  year   = {2024}
}

Comments

32 pages, 1 figure

R2 v1 2026-06-28T15:32:52.968Z