Section: Partnerships and Cooperations

International Initiatives

• Title: Monte Carlo and Quasi- Monte Carlo for rare event simulation

• International Partner (Institution - Laboratory - Researcher):

  • University of Montreal (Canada)

• Duration: started in 2013

• See also: http://www.irisa.fr/dionysos/pages_perso/tuffin/MOCQUASIN/

• The goal of this team is to compute integrals, sums or to solve equations or optimization problems by means of Monte Carlo methods, which are statistical tools used when the models have a high complexity (for instance a large dimension). They are unavoidable methods in areas such as finance, electronics, seismology, computer science, engineering, physics, transport, biology, social sciences... Nonetheless, they have the reputation of being slow, i.e. to require a large computational time to reach a given precision. The goal of the project is to work on acceleration techniques, meaning methods allowing to reach the targeted precision in a shorter computational time than with the standard procedure. A typical framework is that of rare event simulation for which getting even only one occurrence of the event could require a too long computing time. In this case, there are two main acceleration techniques: importance sampling and splitting, on which we work.

• DAT

• Title: Dependability Analysis Tool

• International Partners:

  • Prof. H. Cancela, Univ. of the Republic, Computer Science, Uruguay;

  • Prof. R. Vallejos, UTFSM, Valparaíso, Electrical Eng., Chile;

  • G. Rubino, Dionysos, Inria, general responsible for the project.

• Duration: 2015 - 2016

• Start year: 2015

• The main scientific objective of this project is to develop new techniques to assess the most important dependability properties of a complex system subject to the failures and possible repairs of its components. The central argument behind our proposal is our previous work in the area and some unpublished preliminary and promising results that we believe deserve deep exploration and that should lead to faster evaluation procedures than those available today. We also intend to implement these techniques in an integrated software package usable both in industry and for teaching purposes. Concerning applications, again based on the skills of the participating teams and our past common work, we will illustrate our findings on problems coming from the wireless and optical networking domains.

• SM-HCD-HDD

• Title: Statistical methods for highly complex and/or high dimensional data

• International Partners:

  • Prof. Ricardo Fraiman, Mathematics, Univ. of the Republic, Uruguay, head of the project;

  • many partners in Uruguay, Argentina and Chile,

  • G. Rubino for Dionysos, Inria

• Duration: 2016 - 2017

• Start year: 2016

• The aim of this project is to develop theoretical and computational tools to solve statistical problems that occur when data leaves in high dimensional spaces and/or lives in complex spaces that induce complex geometries.