Emilie OUDOT |
|||||||||||
|
Post-Doc | PhD Thesis | Publications | Teaching | Education |
POST-DOC | |
I'm currently doing a post-doc at INRIA within the VASY team. This post-doc is directed by Radu Mateescu. It is funded by the project "EC-MOAN: Scalable modeling and analysis techniques to study emergent cell behavior: Understanding the E. coli stress response". The research work and expected results of the post-doc are the following:
|
PhD THESIS | ||||||||||||||
I did my PhD thesis at the Laboratoire d'Informatique de l'Université de Franche-Comté (LIFC) within the team Techniques Formelles et à Contraintes (TFC).
My PhD thesis subject was part of the thematic Verification of Timed Systems using Refinement of the TFC team. It was co-directed by Hassan Mountassir and Jacques Julliand, with the participation of Françoise Bellegarde. This thesis was defended on December, 7th 2006 in front of the committee: | ||||||||||||||
| ||||||||||||||
In this work, we focused in particular on timed systems modeled as timed automata in a compositional framework. The aim was to study how to take advantage of the modeling step of a system to cope with the complexity of a verification by model-checking. Indeed, it is well-known that this verification method has the drawback of being difficult to apply on large-sized models. A way out is to check the properties on an abstraction of the model, where the number of states is reasonable enough to allow the verification to be run to completion. Instead of computing an abstraction of the entire model, an alternative is to take advantage of the modeling process of the system, by modeling it incrementally.
For component-based models, incremental modeling can be achieved either by refinement or by integration of components. Both approaches allow to deal with the model-checking problem, by checking properties on smaller models. In particular, refinement allows to handle the verification of global properties (i.e., properties concerning the complete composite model), by checking them on an abstract description of the composite model, instead of on its concrete description. Obviously, established properties have to be preserved by the refinement relation, so that they also hold on the concrete composite model. Integration of components is more adapted to local properties of the components, since the goal is to perform the verification of the properties only on the component they concern. Here again, properties established on a component have to be preserved when the component is merged into an environment. Among others, the main objectives and results of this thesis were the following:
|
PUBLICATIONS |
International Conferences and Workshops with Program Committee and Proceedings |
|
National Conferences with Program Committee and Proceedings |
|
Tool presentations with Proceedings |
|
Others with Program Comittee and Proceedings |
|
Thesis |
|
Internal Reports |
|
Submitted papers |
|
EDUCATION (in french) | ||||||||||||||||||||||||||||
|
Last update on 12/12/2007. |