Majorization, doubly stochastic matrices, and comparison of eigenvalues (Q1120640): Difference between revisions

Given \(x=(x_ i)\in {\mathbb{R}}^ n\), let \(x^.=(x_ i^.)\) denote the decreasing rearrangement of x: \(x_ 1^.\geq x^._ 2\geq...\geq x^._ n\). One says that \(y=(y_ i)\in {\mathbb{R}}^ n\) majorizes x if \(\sum^{k}_{i=1}x^._ i\leq \sum^{k}_{i=1}y^._ i\) for \(k=1,...,n-1\) while \(\sum^{n}_{i=1}x^._ i=\sum^{n}_{i=1}y^._ i.\) A discussion is given showing relationships between majorization and such classical topics as the existence of a doubly stochastic matrix between two matrices, the diagonal equivalence of a matrix to a doubly stochastic matrix, Birkhoff's theorem on doubly stochastic extreme points, portions of the Perron-Frobenius theorem, and others. Egorychev's solution to van der Waerden's problem on permanents, several inequalities concerning determinants, and comparison of the eigenvalues and singular values of a matrix are considered. An elaborate bibliography is cited.