903 A Posteriori Estimation and Adaptive Control of Discretization and Modeling Errors in Computational Sciences and Engineering
Serge Prudhomme, Ecole Polytechnique de Montréal
Ludovic Chamoin, Ecole Normale Supérieure de Cachan
Jens Lang, Darmstadt University of Technology
Advances in computational science and engineering have allowed scientists to contemplate simulations that involve increasingly complex multiphysics and multiscale problems. However, it is also becoming important, perhaps more than ever, to assess the accuracy of the predictions and design suitable adaptive strategies by relying on a posteriori error estimates.
The topic of error estimation and adaptation, globally referred to as model verification, goes now beyond classical discretization error assessment and mesh refinement. It also encompasses adaptive modeling, whose main objective is to adaptively enrich surrogate models derived from model reduction techniques. It further involves novel topics relevant to engineering applications, such as goal-oriented procedures, the computation of guaranteed (upper and lower) error bounds for a large class of physical problems, or the control of errors due to the modeling of uncertainty.
Objectives of the mini-symposium will be to present fundamental contributions to error estimation and adaptive methods and present recent advances in the development of effective methods for the control of approximation errors.
We anticipate contributions on the following topics:
- Estimation of discretization and modeling errors;
- Stability, convergence, and optimality analysis of adaptive methods;
- Computability, practicality, and reliability of error estimates;
- Goal-oriented and adjoint/duality-based techniques;
- Hierarchical, reduced-order, and multiscale modeling;
- Error estimation and adaptive schemes for uncertainty quantification;
- Applications of methods to linear, nonlinear, coupled, or time-dependent problems.
The topic of error estimation and adaptation, globally referred to as model verification, goes now beyond classical discretization error assessment and mesh refinement. It also encompasses adaptive modeling, whose main objective is to adaptively enrich surrogate models derived from model reduction techniques. It further involves novel topics relevant to engineering applications, such as goal-oriented procedures, the computation of guaranteed (upper and lower) error bounds for a large class of physical problems, or the control of errors due to the modeling of uncertainty.
Objectives of the mini-symposium will be to present fundamental contributions to error estimation and adaptive methods and present recent advances in the development of effective methods for the control of approximation errors.
We anticipate contributions on the following topics:
- Estimation of discretization and modeling errors;
- Stability, convergence, and optimality analysis of adaptive methods;
- Computability, practicality, and reliability of error estimates;
- Goal-oriented and adjoint/duality-based techniques;
- Hierarchical, reduced-order, and multiscale modeling;
- Error estimation and adaptive schemes for uncertainty quantification;
- Applications of methods to linear, nonlinear, coupled, or time-dependent problems.
