KOM 518E - Model Predictive Control

Course Objectives

1. To provide the student with an understanding of the fundamental principles in predictive control
To teach to students how to;
2. Solve finite horizon optimal control problems by formulating them as optimization problems.
3. Translate a practical case study into a mathematically well-defined optimal control problem.
4. Construct a range of finite horizon optimal control problems with constraints.
5. Design predictive controllers with guarantees of stability.
6. Use Matlab MPC toolbox and/or write a Matlab program that solves a suitably-defined optimal control problem
7. To provide the student the background information to read research and applications papers on predictive control.

Course Description

Historical Development, Review of optimal control and filtering for linear systems, Review of optimization methods (quadratic and linear programming), Prediction models, Model Predictive Control (MPC) strategy, Classical Model Predictive Control techniques and unconstraint solutions; Dynamic Matrix Control, Model Algoritmic Control, Generalized Predictive Control, Industrial Generalized Predictive Control, Predictive Functional Control, Constraint handling, Solutions for Constraint Problems, Multivariable-Model Predictive Control,Closed loop stability properties of Model Predictive Control systems, Dual Control, Robust Model Predictive Control