Fast and backward stable computation of roots of polynomials, Part II: backward error analysis; companion matrix and companion pencil
Universidad Autónoma de Madrid（马德里自治大学）；Nazarbayev University（纳扎尔巴耶夫大学）；Istituto di Scienza e Tecnologie dell Informazione ‘A. Faedo’ (ISTI), CNR（意大利国家研究委员会信息科学与技术研究所（ISTi））；University of Leuven, KU Leuven（鲁汶大学）；Washington State University（华盛顿州立大学）
Jared L. Aurentz, Thomas Mach, Leonardo Robol, Raf Vandebril, David S. Watkins
本文是J.L. Aurentz, T. Mach, R. Vandebril和D.S. Watkins于2015年发表在《SIAM Journal on Matrix Analysis and Applications》上的论文《快速且具有反向稳定性的多项式根计算》的延续。在该论文中，我们引入了一种伴随QR算法，通过计算伴随矩阵的特征值在$O(n^{2})$时间内找到多项式的根，使用$O(n)$内存。我们证明了该方法是反向稳定的。在此，我们介绍了一种替代方法，即伴随QZ算法，用于解决伴随墨盒的广义特征值问题。更重要的是，我们提供了改进的反向误差分析，利用了问题的特殊结构。改进也部分归因于对转折操作的精度（在理论和实践中）的改进，这是我们算法的关键组成部分。我们证明，对于伴随QR算法，多项式系数的反向误差与系数向量的范数成线性关系。因此，伴随QR算法的反向误差比无结构QR算法（如MATLAB的\texttt{roots}命令）更小，后者系数向量的范数的反向误差增长呈二次关系。伴随QZ算法在适当缩放多项式系数的情况下具有相同的有利反向误差。
英文摘要：
This work is a continuation of "Fast and backward stable computation of roots of polynomials" by J.L. Aurentz, T. Mach, R. Vandebril, and D.S. Watkins, SIAM Journal on Matrix Analysis and Applications, 36(3): 942--973, 2015. In that paper we introduced a companion QR algorithm that finds the roots of a polynomial by computing the eigenvalues of the companion matrix in $O(n^{2})$ time using $O(n)$ memory. We proved that the method is backward stable. Here we introduce, as an alternative, a companion QZ algorithm that solves a generalized eigenvalue problem for a companion pencil. More importantly, we provide an improved backward error analysis that takes advantage of the special structure of the problem. The improvement is also due, in part, to an improvement in the accuracy (in both theory and practice) of the turnover operation, which is the key component of our algorithms.
We prove that for the companion QR algorithm, the backward error on the polynomial coefficients varies linearly with the norm of the polynomial's vector of coefficients. Thus the companion QR algorithm has a smaller backward error than the unstructured QR algorithm (used by MATLAB's \texttt{roots} command, for example), for which the backward error on the polynomial coefficients grows quadratically with the norm of the coefficient vector. The companion QZ algorithm has the same favorable backward error as companion QR, provided that the polynomial coefficients are properly scaled.
XLIM–DMI, UMR CNRS 7252, Faculté des Sciences et Techniques（XLIM-DMI，UMR CNRS 7252，理学院）；School of Mathematical Sciences, Raymond and Beverly Sackler Faculty of Exact Sciences, Tel-Aviv University（特拉维夫大学雷蒙德和贝弗利萨克勒精确科学学院数学系）；Dipartimento di Informatica, Università di Pisa（比萨大学计算机系）
Fast and backward stable computation of the eigenvalues and eigenvectors of matrix polynomials
Instituto de Ciencias Matemáticas, Universidad Autónoma de Madrid（马德里自治大学数学科学研究所）；Nazarbayev University（纳扎尔巴耶夫大学）；ISTI, Area della ricerca CNR（意大利国家研究委员会ISTI研究所）；KU Leuven（荷语鲁汶大学）；Washington State University（华盛顿州立大学）
We study the backward stability of running a backward stable eigenstructure solver on a pencil $S(λ)$ that is a strong linearization of a rational matrix $R(λ)$ expressed in the form $R(λ)=D(λ)+ C(λI_\ell-A)^{-1}B$, where $D(λ)$ is a polynomial matrix and $C(λI_\ell-A)^{-1}B$ is a minimal state-space realization. We consider the family of block Kronecker linearizations of $R(λ)$, which are highly structured pencils. Backward stable eigenstructure solvers applied to $S(λ)$ will compute the exact eigenstructure of a perturbed pencil $\widehat S(λ):=S(λ)+Δ_S(λ)$ and the special structure of $S(λ)…
When computing the eigenstructure of matrix pencils associated with the passivity analysis of perturbed port-Hamiltonian descriptor system using a structured generalized eigenvalue method, one should make sure that the computed spectrum satisfies the symmetries that corresponds to this structure and the underlying physical system. We perform a backward error analysis and show that for matrix pencils associated with port-Hamiltonian descriptor systems and a given computed eigenstructure with the correct symmetry structure there always exists a nearby port-Hamiltonian descriptor system with exac…
A unified approach to Fiedler-like pencils via strong block minimal bases pencils
University of California, Santa Barbara（加州大学圣塔芭芭拉分校）；Universidad Carlos III de Madrid（马德里卡洛斯三世大学）；KU Leuven（荷语鲁汶大学）；Dos Pueblos High School（多斯普埃洛高中）；University of Virginia（弗吉尼亚大学）
Conditioning and backward error of block-symmetric block-tridiagonal linearizations of matrix polynomials
University of California, Santa Barbara（加州大学圣塔芭芭拉分校）；Universidad Carlos III de Madrid（马德里卡洛斯三世大学）；Universidade de Lisboa（里斯本大学）；Faculdade de Economia do Porto（波尔图经济与管理学院）；Boston University（波士顿大学）
A standard approach to compute the roots of a univariate polynomial is to compute the eigenvalues of an associated \emph{confederate} matrix instead, such as, for instance the companion or comrade matrix. The eigenvalues of the confederate matrix can be computed by Francis's QR algorithm. Unfortunately, even though the QR algorithm is provably backward stable, mapping the errors back to the original polynomial coefficients can still lead to huge errors. However, the latter statement assumes the use of a non-structure exploiting QR algorithm. In [J. Aurentz et al., Fast and backward stable comp…
Pole-swapping algorithms, which are generalizations of the QZ algorithm for the generalized eigenvalue problem, are studied. A new modular (and therefore more flexible) convergence theory that applies to all pole-swapping algorithms is developed. A key component of all such algorithms is a procedure that swaps two adjacent eigenvalues in a triangular pencil. An improved swapping routine is developed, and its superiority over existing methods is demonstrated by a backward error analysis and numerical tests. The modularity of the new convergence theory and the generality of the pole-swapping app…
Linearization is a standard method in the computation of eigenvalues and eigenvectors of matrix polynomials. In the last decade a variety of linearization methods have been developed in order to deal with algebraic structures and in order to construct efficient numerical methods. An important source of linearizations for matrix polynomials are the so called Fiedler pencils, which are generalizations of the Frobenius companion form and these linearizations have been extended to regular rational matrix function which is the transfer function of LTI State-space system in [1, 6]. We consider a mul…
We adapt Victor Y. Pan's root-based algorithm for finding approximate GCD to the case where the polynomials are expressed in Bernstein bases. We use the numerically stable companion pencil of Gudbjörn F. Jónsson to compute the roots, and the Hopcroft-Karp bipartite matching method to find the degree of the approximate GCD. We offer some refinements to improve the process.
We consider the uniqueness of solution (i.e., nonsingularity) of systems of $r$ generalized Sylvester and $\star$-Sylvester equations with $n\times n$ coefficients. After several reductions, we show that it is sufficient to analyze periodic systems having, at most, one generalized $\star$-Sylvester equation. We provide characterizations for the nonsingularity in terms of spectral properties of either matrix pencils or formal matrix products, both constructed from the coefficients of the system. The proposed approach uses the periodic Schur decomposition, and leads to a backward stable $O(n^3r)…
A new measure called min-max elementwise backward error is introduced for approximate roots of scalar polynomials $p(z)$. Compared with the elementwise relative backward error, this new measure allows for larger relative perturbations on the coefficients of $p(z)$ that do not participate much in the overall backward error. By how much these coefficients can be perturbed is determined via an associated max-times polynomial and its tropical roots. An algorithm is designed for computing the roots of $p(z)$. It uses a companion linearization $C(z) = A-zB$ of $p(z)$ to which we added an extra zero …
A number of theoretical and computational problems for matrix polynomials are solved by passing to linearizations. Therefore a perturbation theory results for linearizations need to be related back to matrix polynomials. In this paper we present an algorithm that finds which perturbation of matrix coefficients of a matrix polynomial corresponds to a given perturbation of the entire linearization pencil. Moreover we find transformation matrices that, via strict equivalence, transform a perturbation of the linearization to the linearization of a perturbed polynomial. For simplicity, we present t…
One useful standard method to compute eigenvalues of matrix polynomials ${\bf P}(z) \in \mathbb{C}^{n\times n}[z]$ of degree at most $\ell$ in $z$ (denoted of grade $\ell$, for short) is to first transform ${\bf P}(z)$ to an equivalent linear matrix polynomial ${\bf L}(z)=z{\bf B}-{\bf A}$, called a companion pencil, where ${\bf A}$ and ${\bf B}$ are usually of larger dimension than ${\bf P}(z)$ but ${\bf L}(z)$ is now only of grade $1$ in $z$. The eigenvalues and eigenvectors of ${\bf L}(z)$ can be computed numerically by, for instance, the QZ algorithm. The eigenvectors of ${\bf P}(z)$, incl…
Some fast algorithms for computing the eigenvalues of a block companion matrix $A = U + XY^H$, where $U\in \mathbb C^{n\times n}$ is unitary block circulant and $X, Y \in\mathbb{C}^{n \times k}$, have recently appeared in the literature. Most of these algorithms rely on the decomposition of $A$ as product of scalar companion matrices which turns into a factored representation of the Hessenberg reduction of $A$. In this paper we generalize the approach to encompass Hessenberg matrices of the form $A=U + XY^H$ where $U$ is a general unitary matrix. A remarkable case is $U$ unitary diagonal which…
