Chandrasekhar equations for infinite dimensional systems. II: Unbounded input and output case (Q1107090): Difference between revisions

[For part I, see SIAM J. Control Optimization 25, 596-611 (1987; Zbl 0617.49012).] The authors consider the finite-time linear quadratic problem for a linear time-invariant system on a Hilbert space. The system is governed by a \(C_ 0\)-semigroup and unbounded input and output operators. However, the map from the input trajectory to the state is well defined and so is the map from the state to the output trajectory. Under these well posedness conditions they can prove the uniqueness of the optimal trajectory, and the closed-loop system is well defined. The optimal input trajectory u(.) is given by \(u(t)=-B'\Pi (t)U(t,t_ 0)z\), where z is the initial state at time \(t_ 0\), \(u(t,t_ 0)z\) is the state at time t, and \(\Pi\) (t) satisfies the Riccati equation. Furthermore, B'\(\Pi\) (t) satisfies the Chandrasekhar equation. Some of the results in this paper can also be found in a paper by \textit{A. J. Pritchard} and \textit{D. Salamon} [SIAM J. Control Optimization 25, 121- 144 (1987; Zbl 0615.93039)].