The polynomial method for random matrices (Q1029549)

A class of ``algebraic'' random matrices is defined as matrices for which the Stieltjes transform of the limiting eigenvalue distribution function is algebraic, i.e., it satisfies a bivariate polynomial equation. The algebraic random matrices form an important subclass of analytically tractable random matrices and can be effectively studied using combinatorial and analytical techniques that are described in this paper. By defining the class of algebraic random matrices it is possible to extend infinite random matrix theory beyond the special cases of matrices with Gaussian entries. This extension is based on the fact that by encoding probability densities as solutions of bivariate polynomial equations it is possible to use powerful symbolic and numerical technique to compute these densities and their associated moments. In the paper various transform representations are introduced. Techniques for determining the density function of the limiting eigenvalue distribution function and the associated moments are discussed. Also the relevance of the polynomial method to computational free probability is analyzed. Along with the theory of algebraic random matrices a software realization that maps the entire catalog of transformations into symbolic MATLAB code is developed.