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.


Explicit Model Predictive Control for Buffer Management Problems


d a

Semester Thesis, FS15 (10444)

Explicit model predictive control (eMPC) is a method for real-time optimization that circumvents the high online computational demand of usual MPC setups. eMPC allows to precompute optimal control law offline by solving a multi-parametric QP (mpQP) for a chosen set of problem parameters. The result is a piecewise affine (PWA) function. Hence the online computation effort is restricted to a table-lookup, since the optimal input is obtained just by evaluating the PWA control function at each sample. The complexity of the PWA control function varies with the size of the mathematical program (number of parameters, constraints and horizon length). In this semester project we investigate the possibility of using explicit model predictive control (eMPC) for so-called buffer management problems which may represent various dynamic systems. In particular their use was motivated by a case-study on real-time MPC for variable speed drives (VSD). The system which is analysed in this project mimics the dynamics of a VSD in a simplified way.

The mathematical program for eMPC was formulated and an appropriate solver was chosen. Furthermore the possibility to handle time varying constraints with explicit MPC and a numerical study on solution complexity in dependence of the problem size are presented.

Supervisors: Felix Rey, John Lygeros


Type of Publication:

(13)Semester/Bachelor Thesis

J. Lygeros

File Download:

Request a copy of this publication.
(Uses JavaScript)
% Autogenerated BibTeX entry
@PhdThesis { Xxx:2015:IFA_5333
Permanent link