Edge universality of sparse Erd\H{o}s-R\'enyi digraphs

Publication date 10 April 2023

Random graphs (graph-theoretic aspects) (05C80) Random matrices (probabilistic aspects) (60B20)

Abstract: Let

m a t h c a l A

be the adjacency matrix of the ErdH{o}s-R'{e}nyi directed graph

m a t h s c r G (N, p)

. We denote the eigenvalues of

m a t h c a l A

by

l a m b d a_{1}^{c a l A}, ..., l a m b d a_{N}^{c a l A}

, and

| l a m b d a_{1}^{c a l A} | = m a x_{i} | l a m b d a_{i}^{c a l A} |

. For

N^{- 1 + o (1)} l e q p l e q 1 / 2

, we show that [ max_{i=2,3,...,N} �igg|frac{lambda_i^{mathcal A}}{sqrt{Np(1-p)}}�igg| =1+O(N^{-1/2+o(1)}) ] with very high probability. In addition, we prove that near the unit circle, the local eigenvalue statistics of

m a t h c a l A / s q r t N p (1 - p)

coincide with those of the real Ginibre ensemble. As a by-product, we also show that all non-trivial eigenvectors of

m a t h c a l A

are completely delocalized. For Hermitian random matrices, it is known that the edge statistics are sensitive to the sparsity: in the very sparse regime, one needs to remove many noise random variables (which affect both the mean and the fluctuation) to recover the Tracy-Widom distribution. Our results imply that, compared to their analogues in the Hermitian case, the edge statistics of non-Hermitian sparse random matrices are more robust.

This page was built for publication: Edge universality of sparse Erd\H{o}s-R\'enyi digraphs

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6432659)