On the existence of optimal controls for a class of nonlinear infinite dimensional systems (Q1177311)

The paper deals with the existence of optimal controls for strongly nonlinear distributed parameter systems. Specifically the problem under consideration is of the form: \(\dot x(t)\in\partial\psi(x(t))+f(t,x(t))u(t)\) with the state \(x(\cdot)\) belonging to some Hilbert space, \(\partial\psi(\cdot)\) the set of subgradients of a function \(\psi\) and \(u(\cdot)\) the control satisfying \(u(t)\in U(t,x(t))\), specifying a state dependent constraint on the controls. Together with these dynamics, and the initial state \(x(0)=x_ 0\), a cost-index \(J(x,u)=\int^ b_ 0L(t,x(t),u(t))dt\) has to be minimized. It is shown that under appropriate conditions on the system dynamics and cost function \(L\) there exists an optimal pair of state- control functions \((\bar x(t),\bar u(t))\) for the problem. The setting given, allows for an interesting extension to a time optimal control problem with moving target. An example of a strongly nonlinear optimal control problem of parabolic type completes the paper.