Fast Matrix Polynomial Computations
نویسنده
چکیده
Pour une matrice polynomiale P(x) de degr e d dans Mn;n(Kx]), o u K est un corps commutatif, une r eduction a la forme normale d'Hermite peut ^ etre calcul ee en O ~ (M(nd)) op erations arithm etiques, si M(n) est le co^ ut du produit de deux matrices n n sur K. De plus, une telle r eduction peut ^ etre obtenue en temps parall ele O(log +1 (nd)) en utilisant O(L(nd)) processeurs, si le probl eme de d eterminer le premier sous-ensemble ind ependant de l'ensemble des colonnes d'une matrice nd nd sur K, peut ^ etre r esolu en temps O(log (nd)) avec le m^ eme nombre de processeurs. Ces r esultats sont obtenus en g en eralisant au cas matriciel certaines techniques util-is ees pour le pgcd de polyn^ omes scalaires. Mots clefs : algorithmes parall eles, NC 2 K , forme normale d'Hermite, th eorie des r ealisations. Abstract. For a polynomial matrix P(x) of degree d in Mn;n(Kx]) where K is a commutative eld, a reduction to the Hermite normal form can be computed in O ~ (M(nd)) arithmetic operations if M(n) is the time required to multiply two n n matrices over K. Further, a reduction can be computed using O(log +1 (nd)) parallel arithmetic steps and O(L(nd)) processors if the same processor bound holds with time O(log (nd)) for determining the lexicographically rst maximal linearly independent subset of the set of the columns of an nd nd matrix over K. These results are obtained by applying in the matrix case, the techniques used in the scalar case of the gcd of polynomials.
منابع مشابه
Fast and Scalable Parallel Matrix Computations with Optical Buses
We present fast and highly scalable parallel computations for a number of important and fundamental matrix problems on linear arrays with recon gurable pipelined optical bus systems. These problems include computing the Nth power, the inverse, the characteristic polynomial, the determinant, the rank, and an LUand a QR-factorization of a matrix, and solving linear systems of equations. These com...
متن کاملFast and Scalable Parallel Matrix
We present fast and highly scalable parallel computations for a number of important and fundamental matrix problems on linear arrays with reconngurable pipelined optical bus systems. These problems include computing the N th power, the inverse, the characteristic polynomial, the determinant, the rank, and an LU-and a QR-factorization of a matrix, and solving linear systems of equations. These c...
متن کاملDetermination of a Matrix Function in the Form of f(A)=g(q(A)) Where g(x) Is a Transcendental Function and q(x) Is a Polynomial Function of Large Degree Using the Minimal Polynomial
Matrix functions are used in many areas of linear algebra and arise in numerical applications in science and engineering. In this paper, we introduce an effective approach for determining matrix function f(A)=g(q(A)) of a square matrix A, where q is a polynomial function from a degree of m and also function g can be a transcendental function. Computing a matrix function f(A) will be time- consu...
متن کاملFast Error-free Algorithms for Polynomial Matrix Computations
Matrices of pol nomials over rings and fields provide a unifying framework $r many control system design problems. These include dynamic compensator design, infinite dimensional systems, controllers for nonlinear systems, and even controllers for discrete event s stems. An important obstacle for utilizing these owerful matiematical tools in practical applications has been &e non-availability of...
متن کاملAsymptotically fast polynomial matrix algorithms for multivariable systems
We present the asymptotically fastest known algorithms for some basic problems on univariate polynomial matrices: rank, nullspace, determinant, generic inverse, reduced form [8, 9, 16, 17]. We show that they essentially can be reduced to two computer algebra techniques, minimal basis computations and matrix fraction expansion/reconstruction, and to polynomial matrix multiplication. Such reducti...
متن کاملFast approximate computations with Cauchy matrices and polynomials
Multipoint polynomial evaluation and interpolation are fundamental for modern algebraic and numerical computing. The known algorithms solve both problems over any field of constants in nearly linear arithmetic time, but the cost grew to quadratic for numerical solution. We fix this discrepancy: our new numerical algorithms run in nearly linear arithmetic time. At first we restate our goals as t...
متن کامل