论文部分内容阅读
We present nearly O(n)complexity divide-and-conquer methods for finding all the eigenvalues and eigenvectors of a class of symmetric matrices,as well the perturbation analysis.The matrices have certain rank structures,as often encountered in practical applications such as Toeplitz matrices and some discretized problems.We show how to quickly and stably perform the major operations.Eigenvalue approximation accuracies,clustered eigenvalues,and generalizations to SVDs are studied.This is joint work with James Vogel.