English

Multivariate normal approximation for traces of orthogonal and symplectic matrices

Probability 2021-03-08 v1

Abstract

We show that the distance in total variation between (Tr U,12Tr U2,,1mTr Um)(\mathrm{Tr}\ U, \frac{1}{\sqrt{2}}\mathrm{Tr}\ U^2, \cdots, \frac{1}{\sqrt{m}}\mathrm{Tr}\ U^m) and a real Gaussian vector, where UU is a Haar distributed orthogonal or symplectic matrix of size 2n2n or 2n+12n+1, is bounded by Γ(2nm+1)12\Gamma(2\frac{n}{m}+1)^{-\frac{1}{2}} times a correction. The correction term is explicit and holds for all nm4n\geq m^4, for mm sufficiently large. For nm3n\geq m^3 we obtain the bound (nm)cnm(\frac{n}{m})^{-c\sqrt{\frac{n}{m}}} with an explicit constant cc. Our method of proof is based on an identity of Toeplitz+Hankel determinants due to Basor and Ehrhardt, see \cite{BE}, which is also used to compute the joint moments of the traces.

Keywords

Cite

@article{arxiv.2103.03791,
  title  = {Multivariate normal approximation for traces of orthogonal and symplectic matrices},
  author = {Klara Courteaut and Kurt Johansson},
  journal= {arXiv preprint arXiv:2103.03791},
  year   = {2021}
}

Comments

29 pages

R2 v1 2026-06-23T23:48:41.924Z