Bulk universality for generalized Wigner matrices (Q1930864)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Bulk universality for generalized Wigner matrices
scientific article

    Statements

    Bulk universality for generalized Wigner matrices (English)
    0 references
    0 references
    0 references
    0 references
    14 January 2013
    0 references
    In random matrix theory, the dependence of short distance correlation functions for eigenvalues in the bulk of the spectrum on the symmetry class of the ensembles, and their independence on other details of the distributions, is referred to as bulk universality. While the bulk universality has been proven for various classes of invariant ensembles, it is much more difficult to establish it for non-invariant ensembles. Nonetheless, some results -- in particular the local semi-circle law -- have been proven (in particular by some of the authors of this paper) in the last few years for Wigner matrices, with independent and identically distributed (i.i.d.) matrix elements. In this paper, the authors consider bulk universality for some examples of generalized Wigner matrices, whose elements are still independent, but not necessarily identically distributed. More precisely, denoting by \(\sigma^2_{ij}\) the variance of the \((i,j)\) entry of the matrix, the authors consider the ensembles with normalization condition \(\sum_i \sigma_{ij}=1\) for all \(j\). Firstly, under the assumption that \(N\sigma_{ij}^2\) is uniformly bounded from above and below by fixed positive numbers, in the limit \(N\to \infty\) the authors prove the local semi-circle law, down to the smallest possible scale \(N^{-1}\). Secondly, for band matrices, such that \(\sigma_{ij}=0\) for \(|i-j|>M\) for some \(M<N\), the authors establish that the local semicircle law holds to the scale \(M^{-1}\). This is a nicely written paper with important results. Even though it is not that short, the introductory section and the summary of main results in the first few pages are of great help in understanding the achievements and the structure of the paper, and should be of interest even for beginners in this field. Nonetheless, detailed proofs presented in the bulk of the paper require expertise in random matrix theory.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    random matrix theory
    0 references
    random band matrix
    0 references
    semi-circle law
    0 references
    sine kernel
    0 references
    bulk universality
    0 references
    correlation function
    0 references
    eigenvalue
    0 references
    spectrum
    0 references
    Wigner matrices
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references