Random block matrices and matrix orthogonal polynomials (Q975319)

In [\textit{I. Dumitriu} and \textit{A. Edelman}, Ann. Inst. Henri Poincaré, Probab. Stat. 41, No. 6, 1083--1099 (2005; Zbl 1079.15014)], a tridiagonal matrix model was used to study the general beta ensembles. The authors here generalize this tridiagonal matrix model to a so-called random block matrix model, obtained by replacing each entry on or next to the diagonal by a random matrix of fixed size. They then study eigenvalues of these random matrices, and in particular express the asymptotic spectral distribution using Chebyshev matrix polynomials of the first kind. This relation between the theory of random block matrices and the field of matrix orthogonal polynomials seems to be new.