Root-squaring with DPR1 matrices
- 23 Downloads
Recent progress in polynomial root-finding relies on employing the associated companion and generalized companion DPR1 matrices. (“DPR1” stands for “diagonal plus rank-one.”) We propose an algorithm that uses nearly linear arithmetic time to square a DPR1 matrix. Consequently, the algorithm squares the roots of the associated characteristic polynomial. This incorporates the classical techniques of polynomial root-finding by means of root-squaring into a new effective framework. Our approach is distinct from the earlier fast methods for squaring companion matrices. Bibliography: 13 titles.
KeywordsRecent Progress Characteristic Polynomial Fast Method Generalize Companion Classical Technique
Unable to display preview. Download preview PDF.
- 4.J. P. Cardinal, “On two iterative methods for approximating the roots of a polynomial,” in: J. Renegar, M. Shub, and S. Smale (eds.), Proceedings of the AMS-SIAM Summer Seminar “Mathematics of Numerical Analysis: Real Number Algorithms,” Lect. Appl. Math., 32, Amer. Math. Soc., Providence, Rhode Island (1996), pp. 165–188.Google Scholar
- 10.V. Y. Pan, Structured Matrices and Polynomials: Unifed Superfast Algorithms, Birkhäuser/Springer, Boston-New York (2001).Google Scholar
- 13.V. Y. Pan, D. Ivolgin, B. Murphy, R. E. Rosholt, Y. Tang, X. Wang, and X. Yan, “Root-finding with eigen-solving,” in: Dongming Wang and Li-Hong Zhi (eds.), Symbolic-Numeric Computation, Birkhäuser, Basel-Boston (2007), Chap.14, pp. 219–245.Google Scholar