Poles and zeros of matrices of rational functions (Q1177248): Difference between revisions

This is a review paper concerning zeros and poles (including multiplicities) of a rectangular matrix \(G(z)\) of rational functions with coefficients in a field \(F\). The approach is based on the study of various \(F[z]\)-modules associated with \(G(z)\), motivated by ideas of control of linear dynamical systems. In particular, the zero module and the pole module of \(G(z)\) are defined and studied. These modules can be computed using the Smith-McMillan form of \(G(z)\), or, alternatively, using the matrix fractions \(G(z)=N(z)D(z)^{-1}\), where the polynomial matrices \(N(z)\) and \(D(z)\) are relatively coprime. It turns out that, upon counting the dimensions of the zero and pole modules of \(G(z)\), including the zero and pole modules at infinity, the number of zeros of \(G(z)\) is equal to the number of poles of \(G(z)\). The exposition is elementary and requires (beyond linear algebra) only algebra of polynomials and rational functions over a field.