Abstract:
Design of a multivariate control system is a challenging task owing to inherent process nonlinearities, multivariable interactions, and unstable zeros in the process dynamics. Prioritization of multiple control objectives is conventionally achieved using scalarization approach by appropriate weighting factors in the augmented objective function. However, the optimality of these parameters does not hold in presence of unmeasured disturbance and different setpoints. Hence, prioritization in the objectives is not achieved in such circumstances. On the other hand, priority in various objectives can effectively be realized in lexicographic optimization approach. We use the lexicographic optimization approach to address setpoint tracking control problem in linear MPC. To validate the efficacy of the lexicographic ordering approach, multivariate quadruple tank process is considered. Tuning issues with conventional MPC and its remedies using lexicographic approach are discussed in the present work.