Table of Contents
Fetching ...

Global Fluctuations of Gaussian Elliptic Matrices

Lingxuan Wu, Zhi Yin

TL;DR

This work develops a spoke-arc decomposition for non-crossing annular pair partitions to analyze fluctuations in Gaussian elliptic matrices across the ellipticity parameter γ. By isolating spoke-level contributions and factorizing γ through arc weights, the authors obtain closed-form expressions for limiting covariances of trace monomials, valid in both complex and real settings. They show that independent Gaussian elliptic ensembles exhibit asymptotic second-order freeness (complex and real), with the real case incorporating orientation-reversing contributions via the spoke-arc framework. The methodology unifies combinatorial annular structures with arc-weight recursions (Fuss-Catalan connections) to yield quantitative descriptions of second-order fluctuations and their γ-dependence. The results provide a concrete, computable description of covariance structures across the entire ellipticity range and offer a robust tool for analyzing non-normal random matrix ensembles.

Abstract

We introduce a spoke-arc decomposition of non-crossing annular pair partitions $NC_2(p,q)$ that records spoke type and orientation, isolates spoke-level contributions, and factorizes the dependence on the ellipticity parameter $γ$ into a spoke factor and arc weights. This yields closed-form descriptions of the limiting covariance of Gaussian elliptic matrices. As a corollary, we show that an independent family of Gaussian elliptic random matrices is asymptotically second-order free.

Global Fluctuations of Gaussian Elliptic Matrices

TL;DR

This work develops a spoke-arc decomposition for non-crossing annular pair partitions to analyze fluctuations in Gaussian elliptic matrices across the ellipticity parameter γ. By isolating spoke-level contributions and factorizing γ through arc weights, the authors obtain closed-form expressions for limiting covariances of trace monomials, valid in both complex and real settings. They show that independent Gaussian elliptic ensembles exhibit asymptotic second-order freeness (complex and real), with the real case incorporating orientation-reversing contributions via the spoke-arc framework. The methodology unifies combinatorial annular structures with arc-weight recursions (Fuss-Catalan connections) to yield quantitative descriptions of second-order fluctuations and their γ-dependence. The results provide a concrete, computable description of covariance structures across the entire ellipticity range and offer a robust tool for analyzing non-normal random matrix ensembles.

Abstract

We introduce a spoke-arc decomposition of non-crossing annular pair partitions that records spoke type and orientation, isolates spoke-level contributions, and factorizes the dependence on the ellipticity parameter into a spoke factor and arc weights. This yields closed-form descriptions of the limiting covariance of Gaussian elliptic matrices. As a corollary, we show that an independent family of Gaussian elliptic random matrices is asymptotically second-order free.

Paper Structure

This paper contains 17 sections, 15 theorems, 139 equations, 4 figures.

Key Result

Theorem 1.3

Let $X$ be the normalized Gaussian elliptic matrix given in eq:normalized-elliptic, then the limit (as $N \to \infty$) ESD of $X$ is the uniform measure supported in the ellipsoid defined as

Figures (4)

  • Figure 1: Example of a non-crossing annular permutation highlighting cycle orientation. The instance shown is $\pi=(1\,5\,10\,2)(3\,4)(6\,9\,8\,7)\in S_{NC}(4,6)$.
  • Figure 2: An non-crossing annular pair partition $\pi\in NC_2(4,6)$ illustrating the spoke-arc decomposition.
  • Figure 3: Left: a global schematic with inner clusters $A_1,\dots,A_4$ and outer clusters $B_1,\dots,B_4$; spokes realize the order-reversing matching $\sigma$ (here $\sigma(1)=3$, $\sigma(2)=2$, $\sigma(3)=1$, $\sigma(4)=4$). Right: a local diagram of the matched clusters $A_1$ and $B_3$ connected by two spokes; after cutting at the spoke endpoints, each subinterval on the arcs carries a non-crossing pair partition. Here $|I_1|=14$ and $|I_{7}|=18$.
  • Figure 4: Local gluing of connected clusters $I_k$ and $I_{\sigma(k)}$ (two spokes shown, non-crossing pair partitions on arcs suppressed). The endpoints $a,b$ of $I_{\sigma(k)}$ and $c,d$ of $I_k$ are glued pairwise ($a\leftrightarrow c$, $b\leftrightarrow d$) to form a single circle.

Theorems & Definitions (55)

  • Definition 1.1
  • Definition 1.2
  • Theorem 1.3: The elliptic law
  • Theorem 1.4
  • Theorem 1.5
  • Corollary 1.6
  • Theorem 1.7
  • Theorem 1.8
  • Definition 2.1
  • Definition 2.2
  • ...and 45 more