A new method is proposed for designing a multivariable feedback control system. Assuming that the reference inputs to the control system are described by polynomial functions of time, it is possible to determine the final values (steady state values) of the controls which finally make the outputs coincide with the reference inputs. The controls and the outputs at the transient state can be divided into two classes of components: one is that of steady state components and the other the residual variable components. By introducing a quadratic criterion function with regard to only these variable components, a synthesis problem of the feedback control system for the polynomial type reference inputs is reduced to a usual regulator problem. At first the existing condition of the steady state values of the controls for tracking the given reference inputs is derived and then these values are determined. The optimal feedback control system is easily composed by referring the control law which is obtained by solving the regulator problem. This system has the following characteristics:
(1) The reference inputs and their derivatives are applied to the control system.
(2) The parameters of the control system depend only on the parameters of the controlled system and on the weighting matrices of the criterion function.
Some examples are also presented.