A solution algorithm for problems of optimal control in Hilbert spaces (Q1776285)

The paper is concerned with the optimal control problem \[ \min J(u(\cdot))=\int_{t_0}^{\theta} \omega(t,u(t)) dt, \] \[ {\dot x}(t)=Ax(t)+Bu(t),\;t \in I=[t_0,\theta],\;x(t_0)=x_0; \;u(\cdot) \in P, \;Gx(t)+Du(t)=g(t), \;t \in I, \] where \(A\) is a generator of a strongly continuous semigroup of linear operators in a Hilbert space \(B\), \(G\), \(D\) are bounded operators and \(P\) is a convex, closed and bounded set. For solving the problem the authors suggests an iterative algorithm of penalty type and proves its convergence with respect to the functional \(J(\cdot)\) in the presence of noise in the input data. Applications to controlled systems governed by parabolic and functional--differential equations are discussed.