M. Sedraoui
Dichotomy, Tracheotomy_Dichotomy (Trach_Dich), optimization methods, multivariable predictive control
This paper proposes five multidimensional optimization methods used to solve the problem of multivariable predictive control (MPC) under constraints. A distillation column whose dynamics can be represented by an uncertain model is studies. Experiments have proven that performances of the MPC using these suggested methods could be degraded in the case of disturbances acting on the system to be controlled, thus, another original approach is thus proposed. Its principle consists on a combination of two methods: the Dichotomy method and the Tracheotomy method. This new approach is suggested to improve the tracking dynamics of trajectories as well as the rejection of disturbances. The results obtained by simulation with the MPC using these methods compared with those obtained by the MPC using the QP method (Quadratic Programming) available in the Toolbox Optimization of Matlab.
Important Links:
Go Back
