On consistency of linear linearly constrained discrete time systems (Q1066104)

The purpose of this paper is to offer a unitary approach to the problem of consistency conditions for constrained linear discrete-time systems in a state space form. This is made possible by the following proposition: Let m, n be any two positive integers, let P be the nonnegative orthant in \(R^ n\) and let R(.) be the range of a linear operator. Then, given a matrix G(n\(\times m)\) and v in \(R^ n\), the polyhedron P(G,v) is nonvoid if and only if \(v\in R(G)+P\). The set \(R(G)+P\) is called the critical set. In this way the problem of consistency conditions is transfered from the input space to state space and is in the form of membership of a point in the critical set. Therefore, this allows one to connect the study of consistency to concepts as reachability and stability. An important theorem shows that the set of admissible states and the set of the states reachable at t, starting from an admissible state at time 0, are polyhedra.