classical orthogonal polynomials and leverrier-faddeev algorithm for the matrix pencils经典的正交多项式矩阵铅笔和leverrier-faddeev算法.pdfVIP

CLASSICAL ORTHOGONAL POLYNOMIALS AND LEVERRIER-FADDEEV ALGORITHM FOR THE MATRIX PENCILS sE −A ´ ´ JAVIER HERNANDEZ AND FRANCISCO MARCELLAN Received 16 May 2005; Revised 27 March 2006; Accepted 25 April 2006 In this contribution we present an extension of the Leverrier-Faddeev algorithm for the simultaneous computation of the determinant and the adjoint matrix B (s) of a pencil sE −A where E is a singular matrix but det(sE −A) ≡/ 0. Using a previous result by the authors we express B (s) and det(sE −A) in terms of classical orthogonal polynomials. Copyright © 2006 Hindawi Publishing Corporation. All rights reserved. 1. Introduction Consider a linear, time-invariant, multivariable singular system described in the state space as follows: ˙ = Ax + Bu, Ex (1.1) y = Cx, where E ∈ Cn×n is a singular matrix, x is the n-dimensional state vector, u is the m- dimensional input vector, y is the r-dimensional output vector, and A, B, and C are matrices with complex entries and appropriate dimension. We can take the Laplace transform of our system ( 1.1). If det(sE −A) ≡/ 0, then the following transfer function appears: H (s) = C(sE −A)−1B, (1.2) which, in general, is a strictly proper rational matrix (see [ 1, 5] and references therein). The computation of (sE −A)−1 can be carried out by using the Cramer rule, which requires the evaluation of n2 determinants of (n − 1) × (n − 1) polynomial matrices. Clearly, this is not a practical procedure for large n. We will describe an extension of the classical Leverrier-Faddeev algorithm using families of classi