ALADIN has been dissolved on 12/31/2003
It has been replaced by SAGE
It has been replaced by IPSO
Project-Team Presentation
Joint project-team with LABRI (CNRS, University of Bordeaux 1 and ENSEIRB ), located in Bordeaux
Project ALADIN focuses its research activities on the design of numerical schemes and algorithms. Software developed by the project is intended for scientific computing libraries. The work aims to fulfill the criteria of quality, efficiency and reliability.
The execution speed of a numerical algorithm is mainly characterized by its complexity, which is measured by the number of floating point operations, the convergence speed for iterative algorithms and the level of parallelism. Parallel computing is an incentive to look for new algorithms.
The reliability of a numerical algorithm depends on the precision of the approximation (discretization for example), which is often measured by the order of the method, the stability of the resolution scheme and the sensitivity with respect to round-off errors. The search for robust and reliable algorithms is a further, but not exclusive, incentive to look for high performance, parallel algorithms.
Such transversal research has potentially many fields of applications. The project is particularly involved in problems relating to the environment.
Aladin is a joint project of INRIA, CNRS, the University of Rennes 1 and Insa Rennes.
Research themes
- Differential systems:
- Hamiltonian systems,
- Algebraic-differential systems.
- Linear and nonlinear systems of equations:
- Newton-Krylov methods,
- deflation and preconditioning,
- interval computations.
- Eigenvalue problems:
- Arnoldi and Davidson methods,
- proved enclosure of eigenvalues.
International and industrial relations
- Collaboration with Cnes, Cerfacs, Ifremer-Brest, and Simulog.
- Collaboration with the universities of Auckland, Geneva, Minnesota, Patras, Queensland, Sofia, and Yaoundé.
Scientific leader
Jocelyne ERHEL-CHAUX
+33 2 99 84 73 39
Jocelyne.Erhel@inria.fr
Secretary : +33 2 99 84 73 34
