Abstract:
A linear boundary value problem of optimal control is considered in a Hilbert space. The problem is based on linear dynamics and a terminal problem of linear programming at the right end of the time interval. A saddle method is proposed for its solution, and its convergence is proved.