Generalities

• I am the scientific responsible of the research project-team DIONYSOS, which is a major evolution of the former group ARMOR (see the activity reports of 2005 and 2006 for the two last descriptions of ARMOR's activities).
• The main key-words associated with my activities are: perceived quality of multimedia flows over IP networks (QoE analysis), probabilistic modeling of computer and communication systems, analysis of Markovian systems.

Related current responsabilities :

• French head of the associated team PAIR between the university of the Republic in Montevideo, Uruguay, and INRIA.
• Head of the research action RARE on rare event estimation using Monte Carlo.
• Member of IFIP WG 7.3.
• Member of the Steering Board of the Network of Excellence EuroFGI; also responsibli for the Industrial Partnerships in the network (since 2007) and now in the continuation EuroNF.
• Chair of last QEST conference (QEST'08, see the site of the QEST series).
• Chair of RESIM'08, the main workshop on rare event simulation.
  

Some past responsabilities :

• 2003 - 2006: Head of the Scientific Committee of the INRIA Unit at Rennes.
• 1999 - 2002: Responsible of the International Relationships at the INRIA Unit of Rennes.
• 1995 - 1999: Full professor at the ENST Bretagne, responsible for the research at the RSM (Networks and Multimedia Systems) Department.
• 1993 - 2003: Member of the Editorial Board of the International Journal in Operations Research NRL (Naval Research Logistics).
  

1 The PSQA project

I develop a research project called Pseudo-Subjective Quality Assessment (PSQA), which goal is to provide metrics able to automatically quantify the quality of a video or audio flow sent through a packet network such as the Internet, as perceived by the user. The important keywords here are the following:

• automatic: a PSQA metric is evaluated automatically;
• efficient: if necessary, PSQA metrics can be evaluated in real-time;
• perceived quality: PSQA aims at quantifying the quality as perceived by the user;
• accurate: PSQA aims to provide a number that must be close to the number that we would obtain froma controlled subjective experiment performed with real users, as described by appropriate norms depending on the type of media considered.

See the page of the project.

2 Markov Chain Analysis

I am interested in the analysis of Markov chains in both transient and steady-state behaviors. The main applications are in performance or dependability evaluation of computer and communication systems. See the associated page.

3 WAN design

This area concerns basically highly combinatorial optimization problems. The application domain is Wide Area Network (WAN) design. The main tool used is GRASP (Greedy Randomized Adaptive Search Procedure). See the corresponding page.

4 Simulation

We work on rare event simulation in different settings. See the associated page.

4 Misc

See here for some other recent works, in particular on a new paradigm in queueing modeling.