Bulk universality and related properties of Hermitian matrix models (Q2473363)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Bulk universality and related properties of Hermitian matrix models |
scientific article |
Statements
Bulk universality and related properties of Hermitian matrix models (English)
0 references
27 February 2008
0 references
An asymptotic analysis is provided for a class of random matrix ensembles, defined by the probability law \[ P_{n,\beta}(d_{\beta}M)=Z_{n,\beta}^{-1}\exp \biggl\{-\frac{\beta n}{2}\operatorname {tr}V(M) \biggr\}d_{\beta}M.\tag{1} \] where \(M=[M_{jk}]\) is an \(n\times n\) real symmetric \((\beta=1)\) or Hermitian \((\beta=2)\) matrix, the potential \(V:\mathbb{R}\rightarrow\mathbb{R}_+\) is a continuous function and \(Z_{n,\beta}\) is a normalization constant. A new proof is given for the universality properties in the bulk of spectrum of the Hermitian matrix models \((\beta=2)\). The proof is valid under the assumptions that the potential is a globally \(C^2\) and locally \(C^3\) function. This proof is based on the orthogonal polynomial techniques but does not use asymptotics of the orthogonal polynomials. The \textit{sin}-kernel is obtained as a unique solution of a certain nonlinear integro-differential equation that follows from the determinant formula for all marginals of the joint eigenvalue probability density. Motivated by the use in the proof of some facts on the limiting normalized counting measure of the eigenvalues of the matrix models, the authors provide an updated and simplified version of some known results on the existence and the properties of the measure.
0 references
Hermitian matrix models
0 references
universality
0 references
local eigenvalue statistics
0 references
random matrix ensembles
0 references
probability law
0 references
orthogonal polynomial techniques
0 references
nonlinear integro-differential equation
0 references
determinant formula
0 references
eigenvalue probability density
0 references
limiting normalized counting measure
0 references