A regularized Newton solver for linear model predictive control

We investigate direct numerical solvers in linear model predictive control, where the prediction model is given by linear systems subject to linear inequality constraints on the state and the input, and the performance index is convex and quadratic. The inequality constraints are treated by the prim...

Full description

Saved in:
Bibliographic Details
Published in2018 European Control Conference (ECC) pp. 1393 - 1398
Main Authors Malyshev, Alexander, Quirynen, Rien, Knyazev, Andrew, Cairano, Stefano Di
Format Conference Proceeding
LanguageEnglish
Published European Control Association (EUCA) 01.06.2018
Subjects
Complexity theory
IP networks
Linear matrix inequalities
Linear systems
Optimal control
Performance analysis
Predictive control
Online AccessGet full text
DOI10.23919/ECC.2018.8550438

Cover

More Information
Summary:We investigate direct numerical solvers in linear model predictive control, where the prediction model is given by linear systems subject to linear inequality constraints on the state and the input, and the performance index is convex and quadratic. The inequality constraints are treated by the primal-dual interior-point method. We propose a novel direct solver based on the augmented Lagrangian regularization of a reduced Hessian. The new solver has the same arithmetic complexity as the factorized Riccati recursion. The direct solver can be implemented in terms of BLAS3 matrix operations.
DOI:10.23919/ECC.2018.8550438