	Control Group

University of Cambridge > Department of Engineering > Control Group > Publications > Publication

Robust Feasibility in Model Predictive Control: Necessary and Sufficient Conditions

Kerrigan E. C. and Maciejowski J. M.

Proc. 40th IEEE Conference on Decision and Control, December 2001

Abstract

Due to the mismatch between the plant and the model used in an MPC controller, robust constraint satisfaction and feasibility cannot be guaranteed. This paper is concerned with deriving necessary and sufficient conditions for guaranteeing the robust feasibility of an MPC controller wich does not include a model of the disturbance or uncertainty in the on-line prediction. Two practically implementable tests are derived for testing whether or not a given MPC controller is robustly feasible, given an LTI system with bounded state disturbances which is subject to polyhedral constraints on the state and input. The first test assumes that the solution that has been found to the MPC problem is always optimal. The second test determines whether the MPC controller is robustly feasible for all optimal and suboptimal solutions.

Pre-Prints

[PDF]

BibTex Entry

@InProceedings{kerrigan:maciejowski:2001,: author = {Kerrigan E. C. and Maciejowski J. M.},; title = {Robust Feasibility in Model Predictive Control: Necessary and Sufficient Conditions},; address = {Orlando FL, USA},; booktitle = {Proc. 40th IEEE Conference on Decision and Control},; bibkey = {kerrigan:maciejowski:2001},; month = {December},; year = {2001}
}



© 2004 Control Group, Department of Engineering, University of Cambridge www-control-admin@eng.cam.ac.uk