Transfer operators and state spaces for discrete multidimensional linear systems (Q1304667)

This highly complex but interesting paper is best summarized by the authors' abstract: ``This paper treats multidimensional discrete input-output systems from the constructive point of view. We adapt and improve recursive algorithms, derived earlier by E. Zerz and the second author form standard Gröbner basis algorithms, for the solution of the canonical Cauchy problem for linear systems of partial difference equations with constant coefficients on the lattices \({\mathcal N}= \mathbb{N}^{r_1}\times \mathbb{Z}^{r_2}\). These recursive algorithms, in turn, furnish four other solution methods for the initial value problem, namely by transfer operators, by canonical Kalman global state equations, by parametrization of controllable systems and, for systems with proper transfer matrix and left bounded input signals, by convolution with the transfer matrix. In the 2D-case \({\mathcal N}= \mathbb{Z}^2\) the last method was studied by S. Zampieri. Minimally embedded systems are studied and give rise to especially simple Kalman equations. The latter also imply a useful characterization of the characteristics or polar variety of the system by eigenvalue spectra. For \({\mathcal N}= \mathbb{N}^r\) we define reachability of a system and prove that controllability implies reachability, but not conversely. Moreover, we solve, in full generality, the modelling problem which was introduced and partially solved by \textit{F. Pauer} and \textit{S. Zampieri} [J. Symb. Comput. 21, 155-168 (1996; Zbl 0869.39001)]. Various algorithms have been implemented by the first author in axiom, and examples are demonstrated by means of computer generated pictures. Related work on state space representations has been done by the Padovian and Groningian system theory school.'' The introduction of this paper, together with page 21 and 22 of section 3 put the results into perspective, a perspective that certainly is needed to appreciate the contributions of Kleon and Oberst.