Another approach to long-range predictive control

摘要

A long-range predictive control based on a quadratic cost function is developed. The controlled system is described by an input-output model and, in order to compute the control gain matrix via the solution of a discrete-time Riccati equation, a state-space equation is formulated through a partitioned output prediction-block representation of the system. Extension to the infinite horizon case is possible and some stability results can be obtained.