Separable type representations of matrices and fast algorithms. Volume 2. Eigenvalue method (Q1945556)

The volume 2 of the book is concerned with the eigenvalue problem for matrices with quasiseparable structure. It is organized in four parts, including 16 chapters. The first part (Part V of the book; Chapters 21--24) is devoted to quasiseparable of order one matrices. The authors analyze important properties such as: characteristic polynomial, eigenvalues and corresponding eigenspace structure, kernels and multiple eigenvalues. Part two (Part VI of the book; Chapters 25--28) is devoted to divide and conquer type methods for the eigenvalue problem. The first two chapters analyze separately the two aspects -- divide and conquer -- whereas in the last two ones, complete algorithms are described for Hermitian and unitary Hessenberg matrices. The third part (Part VII of the book; Chapters 29--31) deals with the QR iterative algorithm for eigenvalues, the reduction to Hessenberg form algorithm and implicit QR method for upper Hessenberg matrices. In Part four (Part VIII of the book; Chapters 32--36) the authors are concerned with the QR iterative algorithm for companion matrices. They present explicit QR iteration, implicit methods with compression, factorization based methods and implicit algorithms based on QR representation. As a continuation of the first volume, the second one addresses to engineers, scientists, numerical analists, computer scientists and mathematicians alike. For Volume 1 see the review above [(2014; Zbl 1280.65027)].