UWSpace is currently experiencing technical difficulties resulting from its recent migration to a new version of its software. These technical issues are not affecting the submission and browse features of the site. UWaterloo community members may continue submitting items to UWSpace. We apologize for the inconvenience, and are actively working to resolve these technical issues.
 

Computing Matrix Canonical Forms of Ore Polynomials

Thumbnail Image

Files

Khochtali_Mohamed.pdf (390.07 KB)

Date

2017-05-01

Authors

Khochtali, Mohamed

Journal Title

Journal ISSN

Volume Title

Publisher

University of Waterloo

Abstract

We present algorithms to compute canonical forms of matrices of Ore polynomials while controlling intermediate expression swell. Given a square non-singular input matrix of Ore polynomials, we give an extension of the algorithm by Labhalla et al. 1992, to compute the Hermite form. We also give a new fraction-free algorithm to compute the Popov form, accompanied by an implementation and experimental results that compare it to the best known algorithms in the literature. Our algorithm is output-sensitive, with a cost that depends on the orthogonality defect of the input matrix: the sum of the row degrees in the input matrix minus the sum of the row degrees in the Popov form. We also use the recent advances in polynomial matrix computations, including fast inversion and rank profile computation, to describe an algorithm that computes the transformation matrix corresponding to the Popov form.

Description

Keywords

Computer Algebra, Symbolic Computation, Linear Algebra, Popov Form, Hermite Form, Fast Algorithms

LC Keywords

Citation

URI

http://hdl.handle.net/10012/11831

Collections

Theses
Computer Science