3
$\begingroup$

I am interested in explicit expression or bounds for the Fourier transform (characteristic function) of the joint probability distribution of eigenvalues of random matrices $X\sim \mathrm{GUE} (d)$, where $\mathrm{GUE} (d)$ stands for Gaussian Unitary Ensemble in dimension $d$.

The expression for the joint distribution of eigenvalues $=\lambda_1,\ldots,\lambda_d$ of matrices in this ensemble is well-known

$p_{\mathrm{GUE}(d)}(\lambda_1,\ldots,\lambda_d) = N_d \prod_{1\leq i<j\leq d} (\lambda_i-\lambda_j)^2 \exp(-\frac{d}{2}\sum_{i=1}^d \lambda_i^2)$ ,

where $N_d$ is a normalisation constant.

However, nowhere in the literature, I could find information about the Fourier transform of $p_{\mathrm{GUE}(d)}$ i.e

$f_{\mathrm{GUE}(d)}(k_1,\ldots,k_d) = \int_{\mathbb{R}^d}d\lambda_1 \ldots d\lambda_d \exp(i\sum_{j=1}^d k_j \lambda_j )p_{\mathrm{GUE}(d)}(\lambda_1,\ldots,\lambda_d) $ .

This was surprising to me since characteristic functions seem a rather natural object to study and GUE is one of the basic enables appearing in random matrix theory.

I am especially interested in understanding the behaviour (decay) of $f_{\mathrm{GUE}(d)}(k_1,\ldots,k_d) $ as a function of $|k|=\sqrt{\sum_{i=1}^d k_i^2}$ for large $d$.

My motivation to study this problem comes from some considerations at the intersection of quantum chaos and quantum computing (particularly, the problem of "complexity growth" in unitary evolution $\exp(it H)$, where $H\sim\mathrm{GUE}(d)$).

$\endgroup$

2 Answers 2

6
$\begingroup$

Complementing the answer by Carlo, if you take all $k$'s equal you have $$f_{\rm GUE(d)}(k,...,k)\propto \int dX e^{ik{\rm Tr}(X)}e^{-\frac{d}{2}{\rm Tr}(X^2)}.$$ Taking $x$ to be any real diagonal element from $X$, this is $$f_{\rm GUE(d)}(k,...,k)\propto \left(\int dx e^{ikx}e^{-\frac{d}{2}x^2}\right)^d.$$

I think in the end you have simply $f_{\rm GUE(d)}(k,...,k)=e^{-k^2/2}$.

$\endgroup$
3
  • $\begingroup$ you seem to assume that a diagonal element of $X^2$ is the square of a diagonal element of $X$ ... that seems wrong... $\endgroup$ Apr 13, 2021 at 15:01
  • $\begingroup$ @CarloBeenakker No, I just write $e^{-{\rm Tr}(X^2)}=\prod_{ij}e^{-|x_{ij}|^2}$. The integral over the non diagonal elements is trivial $\endgroup$
    – Marcel
    Apr 13, 2021 at 15:11
  • $\begingroup$ right, my mistake. $\endgroup$ Apr 13, 2021 at 15:18
4
$\begingroup$

The Fourier transform of the marginal distribution of a single eigenvalue in the GUE is known, $$f_{{\rm GUE}(d)}(k,0,0,\ldots,0)=e^{-\tfrac{1}{2}k^2/d}\sum_{j=0}^{d-1}(-1)^jk^{2j}\frac{(d-1)(d-2)\cdots(d-j)}{j!(j+1)!d^j},$$ see these lecture notes.

A curiosity: the $d=2$ result is given in this publication in terms of the confluent hypergeometric function $M(-1,1/2,k^2/8)$, without the realization that this is simply $1-k^2/4$.

$\endgroup$
2
  • $\begingroup$ Thanks for the answer. However, I am interested in the Fourier transform of the full distribution, not just the marginals. $\endgroup$ Apr 13, 2021 at 14:34
  • 1
    $\begingroup$ I would be surprised if there is a closed form expression for that, you would have a closed form expression for correlators of arbitrary order... $\endgroup$ Apr 13, 2021 at 14:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.