Joint projet-team with CNRS and the
University of Provence through the Laboratoire d’Analyse,
Topologie et Probabilités (UMR 6632). It is located in the
"Technopôle" of Château Gombert in Marseilles.
This project aims at developing numerical probability methods.
Its main competence fields are stochastic differential
equations, partial differential equations with random
parameters and finally stochastic partial differential
equations.
These objects are studied in parallel with Markov chains in
random medium or not. We also take an interest in stochastic algorithms.
We also take an interest in stochastic algorithms. These tools are applied to different fields.
Research themes
Study of random media. The aim is to study the properties
of random processes (diffusion process, Markov chains) in
a random domain. These objects are also linked to some
partial differential equations with random parameters. For
instance, in reservoir engineering, we study a partial
differential equation deduced from Darcy law where the
permeability field is supposed to be random. This first
axe finds natural applications in reservoir engineering,
in the domain of semiconductors, etc.
Numericals methods for turbulence phenomenon simulation is a wide
range and complex field. We focus on "probabilistic"
modelisation of specific models like two dimensional
transport equation in gaussian, incompressible random
fields.
We keep on applying the results obtained in numerics for
non linear filtering and identification.
Numerical implementation. A very special effort is made
for the implementation of reliable software on different
platforms (massively parallel computers, vectorial
computers, station network, etc.) using adapted languages (parallel Fortran, parallel C language, MPI, PVM, etc.). In these different research fields, some toolboxes are on study.
International and industrial relations
International and industrial relations
International and industrial relations
About the first research theme, an industrial
collaboration is carried out with IFP (Institut Français du Pétrole).
Scientific collaborations:
Random media: University of Provence, Princeton
University, Moscow Academy of Sciences, Ecole
Polytechnique.
Non linear identification: Irisa, University of
Southern California, University of Bordeaux.
Contracts:
Program "modélisation et simulation numérique" du CNRS:
Leader of the group "Diffusion en milieu aléatoire"
