Extremal dynamic errors in linear dynamic systems

Two different analytical methods of determining extremal dynamic errors in linear dynamic systems are presented. The main idea of these methods is based on finding certain additional equations. These additional equations are obtained due to the assumption that an extremal point obtained from the necessary condition dx/dt=0, is also an extremum point with respect to initial conditions, that is, d/dci= 0, i = 1, . . . , n.