Title :
Feedback approximation in model predictive control
Author_Institution :
Saint Petersburg State University, 7/9 Universitetskaya nab., 199034, Russia
Abstract :
Model predictive control of a nonlinear dynamical system with linearizable and Lipschitz right-hand side is studied. It is proven that the exact model predictive control feedback allows approximation with a piecewise-affine, possibly discontinuous, function defined by its precalculated values on a finite grid. Approximation is understood not in the sense that control values are close but rather in the sense of closeness of the cost values. Additionally, if such approximation is close enough, stability of the approximate MPC is proven.
Keywords :
"Predictive control","Power system stability","Approximation algorithms","Stability analysis","Prediction algorithms","Linear approximation"
Conference_Titel :
"Stability and Control Processes" in Memory of V.I. Zubov (SCP), 2015 International Conference
DOI :
10.1109/SCP.2015.7342134