Exponential stability of discrete linear repetitive processes (Q4783546)

A discrete linear repetitive process has the representation \[ \begin{aligned} lx_{k+1}(t+1)&= Ax_{k+1}(t)+Bu_{k+1}(t)+B_{0}y_{k}(t)\\ y_{k+1}(t)&= Cx_{k+1}(t)+Du_{k+1}(t)+D_{0}y_{k}(t),\end{aligned}\tag{1} \] \(k\in\mathbb{Z}^{+}, t\in[0,\alpha]\)= a set of integers; the function \(x:\mathbb{Z}^{+}\times[0,\alpha]\to E\) is the pass state vector, \(y:\mathbb{Z}^{+}\times[0,\alpha]\to W\) is the pass profile vector and \(u:\mathbb{Z}^{+}\times[0,\alpha]\to V\) is the input vector, \(E,V\) and \(W\) are finite-dimensional normed linear spaces over \(\mathbb{C}\) and \(A,B,B_{0},C,D,D_{0}\) are linear operators. This process is studied by using the properties of an associated discrete Volterra operator and its representation in the ring of power series in one indeterminate. A stability theory is developed for the homogeneous version of the system (1) (i.e. with \(u\equiv 0\)), by defining two distinct concepts termed weak exponential (pass) stability and exponential stability. These definitions require that both the state and pass profile vector sequences generated by the system from bounded boundary conditions converge strongly to \(0\) at the same geometric rate as \(k\to\infty\). The essential difference between these definitions is that the former requires this property with respect to the (finite and fixed) pass length \(\alpha\) and the latter requires \(\alpha=\infty\). A sufficient condition for weak exponential stability is given, and it is proved by an example that this condition is not necessary. Exponential stability is completely characterized by a necessary and sufficient condition which shows that this property is equivalent to the stability along the pass property previously developed by three of the authors. It is emphasized that this equivalence can also be viewed as the counterpart of the one (previously established by Pandolfi) between BIBO stability and exponential stability of the 2D Fornasini-Marchesini state space model.