Reduced basis methods applied to obstacle problems and parametrized distributed optimal control problems with control and state constraints

Aachen (2016) [Dissertation / PhD Thesis]

Page(s): 1 Online-Ressource (174 Seiten) : Illustrationen, Diagramme


In this thesis, we address new reduced basis approaches for parametrized variational inequalities and parametrized linear-quadratic optimal control problems with constraints on both the control and the state. These problems are challenging because of their nonlinear nature and, after a discretization, usually large number of degrees of freedom. We address the latter difficulty by reducing the complexity of solving suchproblems by applying the reduced basis method, resulting in very small approximate problems. These small problem cannot only be inexpensively solved but also provide inexpensive error estimators/bounds that allow us to directly control the approximation error.Many models in computation engineering science require numerous evaluations of input-output systems that are described by parameter-dependent partial differential equations. Inputs might determine system properties, e.g., the conductivity coefficients, whereas the system's outputs are then, e.g., the heat distribution. Evaluating these input-output relationships usually requires expensive and iterative computations, which makes the need for rapid and yet reliable reduction methods apparent.



Bader, Eduard


Veroy-Grepl, Karen Paula
Herty, Michael Matthias


  • URN: urn:nbn:de:hbz:82-rwth-2016-093640
  • REPORT NUMBER: RWTH-2016-09364