Department of
Scientific Computations
Research directions
- Numerical Methods for Partial Differential Equations (Finite
Difference Schemes and Finite Element Method): derivation of discrete
problems, a priori estimates and error analysis; fast (direct)
elliptic solvers; approximate wavelet stabilized hierarchical basis
methods; least-squares mixed finite element methods (error analysis
and iterative methods).
- Computational Linear Algebra (Iterative Methods and Algorithms,
Preconditioning, Sparse Matrices): (block-) incomplete factorization
preconditioners; multigrid and domain decomposition methods in
subspace; algebraic multilevel preconditioning methods; iterative
refinement methods of BEPS type; iterative methods for nonsymmetric
discretization problems; hierarchical and multilevel methods for mixed
finite elements and nonconforming elements; preconditioning
saddle-point operators; preconditioners based on direct decompositions
of the discretization spaces.
- Applications of Parallel Algorithms and Supercomputing
(Large-Scale Problems, Parallel or/and Vector Computers, Clusters of
Workstations)
