Functional CLT for non-Hermitian random matrices

exhibits Gaussian fluctuations as the matrix size grows to infinity, which consists of two independent modes corresponding to the tracial and traceless parts of

A

. We find a new formula for the variance of the traceless part that involves the Frobenius norm of

A

and the

L^{2}

-norm of

f

on the boundary of the limiting spectrum.

Review text

Let \(X\) be an \(N\times N\) random matrix whose entries are independent and identically distributed real or complex random variables with zero mean and variance \(1/N\), and let \(A\) be a deterministic and norm-bounded \(N\times N\) matrix. The authors study fluctuations of the statistics \[ \text{Tr}f(X)A = \frac{1}{2\pi\text{i}}\oint_\gamma f(z) \text{Tr}\left((z-X)^{-1}A\right)\text{d}z\,, \] where \(f\) is a function which is analytic in a neighbourhood \(\mathcal{D}\) of the closed unit disk, and \(\gamma\) is a positively oriented curve in \(\mathcal{D}\) which encircles the spectrum \(\sigma(X)\) of \(X\). The authors show that the event where \(\sigma(X)\subset\mathcal{D}\) on which these statistics are well defined has high probability, and that the fluctuations of these statistics are asymptotically Gaussian with constant order as \(N\to\infty\).

Recommendations

Cites work

Cited in

(1)

Multi-point functional central limit theorem for Wigner matrices

This page was built for publication: Functional CLT for non-Hermitian random matrices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6147700)