Technical Report EE04025 - Notes on Linear Model Predictive Control

Linear model predictive control (MPC) assumes a linear system model, that the constraints sets are representable via linear inequalities and that the objective function is convex quadratic. Linear MPC is appealing because the associated optimisation problem typically solved at each time interval may be expressed as a convex quadratic program, which can be solved efficiently online. Topics not covered include methods for solving quadratic programs, soft constraints, terminal state constraints, closed-loop stability, closed-loop robustness and optimal reference trajectory calculation. A Matlab simulation example is included for illustration.