A random matrix decimation procedure relating \(\beta = 2/( r + 1)\) to \(\beta = 2( r + 1)\) (Q2517974): Difference between revisions

The Dixon-Anderson conditional probability density function refers to the function of \(\{\lambda_i\}\) [cf. \textit{T. H. Baker} and \textit{P. J. Forrester}, Commun. Math. Phys. 188, No. 1, 175--216 (1997; Zbl 0903.33010)], where inequalities for \(\lambda_i\) specify an interlaced region. Similarly, an analogous conditional probability density function for angles \(\{\Psi_j\}_{j=1, \dots, n}\) due to the author and \textit{E. M. Rains} [Int. Math. Res. Not. 2006, No.~5, Article ID 48306 (2006; Zbl 1117.15026)] is given with the conditions for angles in the interlaced region. In the paper, different generalizations of mentioned probability density functions are given and their consequences to random matrix theory are discussed. It is shown that they provide inter-relations between matrix ensembles with orthogonal, unitary and symplectic symmetry. These inter-relations when applied on joint eigenvalue probability density functions provide linking between the joint distribution in certain \(\beta\)-ensembles with \(\beta = 2/(r+1)\) and \(\beta = 2(r+1)\) for marginal distributions of single eigenvalues.