A solution to the optimal pursuit problem for distributed parameter systems (Q2526059)

The dynamical system is given by a linear differential equation of the type \[ \dot u = Au + Bf,\quad 0<t<T, \] \[ \text{subject to } \Vert u(t) - u_0\Vert \to 0\text{ as }t\to 0^+, \] \(u_0\in D(A)\) with \(u\in H(T)\), \(f\in F(T)\), where \(H(T)\) and \(F(T)\) denote suitable spaces of square-integrable functions taking their values in appropriate Hilbert spaces \(H\) and \(F\), respectively. \(B\) is a certain bounded linear operator, \(A\) is a certain unbounded linear operator with domain \(D(A)\subset H\), and \(A\) is the infinitesimal generator of a strongly continuous semigroup of linear operators. The optimization problem considered is to minimize a quadratic functional \[ J(f) = \Vert u(f) - u_d\Vert^2 + \lambda^2 \Vert f\Vert^2\quad\text{on }F(T), \] where \(u(f)\) is the response to control \(f\), and \(u_d\) is a given desired state in \(H(T)\). Existence, uniqueness, and sufficiency theorems are established for this problem by means of an associated two-point boundary value problem. An algorithm for generating the optimal control and an example are given.