Several linear and higher order methods to compute nth roots of a given real or complex matrix are presented in this paper. These include Newton-like, subspace, and Krylov type methods. As a special case, the matrix sector function and other roots of an identity matrix are computed and shown to be an efficient numerical tool for computing a block eigendecomposition of a given matrix.
|Original language||English (US)|
|Number of pages||6|
|Journal||Proceedings of the IEEE Conference on Decision and Control|
|State||Published - Dec 1 2001|
|Event||40th IEEE Conference on Decision and Control (CDC) - Orlando, FL, United States|
Duration: Dec 4 2001 → Dec 7 2001