ABSTRACT
The schemes related to model predictive control (MPC) were developed in late 1970s and became known as dynamic matrix control or model algorithmic control. These have found applications in the process industry, and with the availability of fast computers/microprocessors, MPC has found several applications in robotics, automobiles, nuclear, and aerospace industries. One of the major aspects of MPC formulation is the state estimator that is used to carry out online/real time prediction. Sometimes a mathematical model used for developing an industrial MPC strategy is obtained from the operating data, and the states might not have any physical meaning; thus, MPC is formulated as an output control scheme. The MPC scheme is based on minimization of a performance measure, and the resulting formulation should guarantee stable closed loop behavior. The MPC is a regulatory control that uses an explicit dynamic model of the responses of the process variables to compute the control “actions,” also known as “moves”.
