A state space approach to minimum variance control of multivariable ARMAX systems

The minimum variance control problem for multivariable ARMAX systems is studied using a state-space approach. By reformulating it as a singular quadratic control problem, the stabilizing minimum variance control strategy is obtained by considering the stabilizing solution to the associated singular control algebraic Riccati equation. Various useful connections between the ARMAX and state-space representations of the systems are also found. These make it possible to establish equivalence of the state-space minimum variance control law and the ARMAX control law.<
>