In this paper we initiate a program to study the controllability properties of matrix eigenvalue algorithms arising in numerical linear algebra. Our focus is on a well-known eigenvalue method, the inverse power iteration deΓΏned on projective space. A complete characterization of the reachable sets a
On controllability of the real shifted inverse power iteration
β Scribed by Uwe Helmke; Fabian Wirth
- Publisher
- Elsevier Science
- Year
- 2001
- Tongue
- English
- Weight
- 198 KB
- Volume
- 43
- Category
- Article
- ISSN
- 0167-6911
No coin nor oath required. For personal study only.
β¦ Synopsis
Controllability properties of the inverse power method on projective space are investigated. For complex eigenvalue shifts a simple characterization of the reachable sets in terms of invariant subspaces can be obtained. The real case is more complicated and is investigated in this paper. Necessary and su cient conditions for complete controllability are obtained in terms of the solvability of a matrix equation. Partial results on conditions for the solvability of this matrix equation are given.
π SIMILAR VOLUMES
Let z0 be an arbitrary point in the complex plane. For each positive integer n we choose sn(z) to be -5/(1 + z) or -0.5/(1 + z) with equal probability. We introduce the orbit (zn)~, where zn = Sn(Zn-1) for n > 1. We prove that with probability one the orbit is attracted to the real axis. In the proo