Computing generalized inverse of polynomial matrices by interpolation (Q2369112)

Two algorithms for computing the Moore-Penrose inverse of polynomial matrix are compared. The first one is the extension of the Leverrier-Faddeev method, see \textit{N. P. Karampetakis} [Linear Algebra Appl. 252, 35--60 (1997; Zbl 0869.65028)]. The second one uses the polynomial interpolation and is similar to \textit{A. Schuster} and \textit{P. Hippe} [IEEE Trans. Automat. Control 37, No. 3, 363--365 (1992)]. Complexity analysis is made for both algorithms, and they were tested on several classes of test examples. The interpolation algorithm seems to be faster on dense matrices.