Home  |  Chapter  |  News   |   Personalities  |   Service  |   e-Journal  |   History  |  Pantheon  |  Bibliography  Rus / Eng  |   Contacts  |   Forum   |   Links

 

 

Marc Bouissou Professor at Ecole Centrale Paris Research interests: Complex systems dependability: evaluation, optimization, allocation Modeling of complex systems: automatic generation of reliability models (fault-trees, Markov chains, Bayesian networks...) from highly generic and very powerful formalisms (FIGARO language, BDMP (Boolean logic Driven Markov Processes) ®). Extensions of Bayesian networks: probabilistic relational models, object oriented Bayesian networks. Optimization of the processing of the generated models in order to limit combinatorial explosion problems: techniques relying on BDD (Binary Decision Diagrams) for Boolean models: variable ordering heuristics, techniques based on sequence exploration and quantification for Markov models, probabilistic inference algorithms for Bayesian networks, Monte-Carlo simulation. Dependability evaluation of software intensive systems. Programmable systems pose specific problems, because a large part of their failures are due to design errors. This is why models used for conventional systems studies are not well adapted. Bayesian networks, because they are able to model inference on uncertain data, allow to formalize the reasoning of experts in charge of the assessment of such systems. Current research: Object oriented bayesian belief networks Reliability of very high voltage grids e-mail: Marc.Bouissou@edf.fr

Detailed Information  |  Publications  |  Hobbies

© 2006 - International Group On Reliability   |   Send your correspondence to the Website hosts: Igor Ushakov & Alexander Bochkov   |   Updated: 2007, August 18