On a new computing technique in optimal control theory and the maximum principle.