Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Comparison of ADMM Formulations for MPC Problems


D. Manickathu

Master Thesis, HS15 (10420)

In this work we apply the alternating direction method of multipliers (ADMM) toMPC problems with linear system dynamics and box constraints. A reformulation of the MPC problem needs to be done in order to be able to solve it with ADMM. There are a multitude of ways to reformulate an MPC problem to the standard ADMM form. The system dynamics can be expressed by utilizing the sparse or dense stacking method, state elimination can be applied to reduce the system dimension and there are various ways to choose a splitting of the objective function. All these options lead to di↵erent ADMM formulations. This thesis analyses the e↵ect of these reformulations on the convergence rate of the ADMM algorithm. Furthermore the practicability of optimal parameter selection for certain formulations is evaluated. A simulation environment was written in Matlab which generates random MPC problems to compare the various formulations with respect to the convergence speed of ADMM. We also investigate how ADMM performs when we apply certain theoretical results for optimal parameters on suitable formulations.

Supervisors: Felix Rey, John Lygeros


Type of Publication:

(12)Diploma/Master Thesis

J. Lygeros

File Download:

Request a copy of this publication.
(Uses JavaScript)
% Autogenerated BibTeX entry
@PhdThesis { Xxx:2016:IFA_5425
Permanent link