logo sfb mobis
 Mathematical Optimization and Applications in Biomedical Sciences
SFB Research Center 


 Introduction
    Home
    Objectives
    Subprojects:

 People
    Office
    Researchers
    Collaborators
    Jobs

 Events
    Seminars
    Visitors
    Meetings

 Products
    Publications
    Prizes
    Patents
    Test data
    Theses

 Commentaries
    Press

 Internal
    Templates
    empty space





OPTIM
FREELEVEL
FEMBEM
HEART
MRI

Subproject OPTIM

Numerical Methods for Optimal Control with PDE Constraints

Direction: Karl Kunisch
 

Keywords: model reduction, inertial manifold, proper orthogonal decomposition, adaptivity, semi-smooth Newton

Information: The central goal of this project is the advance of analytical and numerical methods for large scale optimal control problems in the context of partial differential equations. It contains method driven and problem driven components alike. Control of partial differential equations has experienced tremendous advances over the last two decades. To have a strong impact on science it must contribute, with increasing vigour, to real world problems. In this respect one of the aims of OPTIM is to strongly collaborate with HEART on deriving optimal, open loop, controls for defibrillation of arrhythmias in the heart.

OPTIM will also derive suboptimal closed loop controls, based on receding horizon techniques. At a later stage closed loop strategies will be based on the Hamilton Jacobi Bellman equation. New methodologies for treating the large scale aspect will be based on model reduction methods. For the first time an analysis of the use of approximate inertial manifold techniques in the context of optimal control will be provided. As an alternative to obtain reduced order models, the method of proper orthogonal decomposition, which was co-initiated by the authors of the proposal, will be significantly advanced by incorporating adaptivity issues related to basis updates, and snapshot location. Semi-smooth Newton methods and related path-following techniques, necessary in cases of low Lagrange multiplier regularity will be investigated for mixed control-state constraints and in the context of time optimal control.

  Med Uni Graz    TU Graz    Uni Graz Webmaster WT / Mar 15, 2016