Section: Partnerships and Cooperations

International Initiatives

In the context of CIRIC, we cooperate with the team of Reinaldo Vallejos, professor at the UTFSM, Valparaíso, Chile, on different topics related to networking and modeling issues. Specifically, these activities are organized around two collaborative projects, 8.3.2.1 and 8.3.4.1 , where one can find the scientific details. We start to work also with Javiera Barrera's team at the University Adolfo Ibañez, Santiago de Chile, on stochastic optimization problems.

• Title: MANAP (Markovian ANalysis and APplications)

• International Partner (Institution - Laboratory - Researcher):

  • Universidad Técnica Federico Santa María (Valparaíso, Chile)

• Duration: 2013 - $\cdots$

• See also: http://people.rennes.inria.fr/Gerardo.Rubino/RESEARCH/MANAP/manap.html

• This project has two goals. The main one is to develop techniques allowing to accelerate solving techniques for Markov models, both in equilibrium and in their transient phases. The applications behind these efforts are the evaluation of performance, dependability and performability properties of complex communication systems. The second goal is to apply these solving techniques to specific problems in networking, concerning optical and wireless infrastructures. On both cases there is some emphasis on dependability aspects: fault tolerance routing schemes in the first case, complex dependability characteristics and their analysis in the second one. A third explicit objective is to produce a software tool implementing the techniques coming from the main research direction of the project, designed to be used both by engineers and researchers.

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

• International Partner (Institution - Laboratory - Researcher):

  • University of Montreal (Canada)

• Duration: 2013 - $\cdots$

• 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 tools 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. A typical framework is that of rare event simulation for which getting even only one occurrence of the event could require a very long time. In this case, there are two main acceleration techniques: importance sampling and splitting, on which we work.

• Program: Stic AmSud

• Title: Accelerating Markov Models for analysis and design of dynamic WDM optical networks (AMMA)

• Inria principal investigator: Gerardo Rubino

• International Partners (Institution - Laboratory - Researcher):

  • University of the Republic (UDELAR), Montevideo, Uruguay – Computer Science at the Engineering Faculty – prof. Héctor Cancela

  • Technical University Federico Santa María (UTFSM), Valparaíso, Chile – Electronics Department – Prof. Reinaldo Vallejos

• Duration: 2 years, Jan. 2013 – Dec. 2014

• This project has two main scientific goals: (i) to develop methods capable of solving Markov models faster than with state-of-the-art techniques, and (ii) to apply these techniques to the design of fault-tolerant optical networks. The rationale behind (i) is that the group has ideas and some preliminary promising unpublished results that makes it expect that its approach will be effective in producing new nice solving procedures. Concerning (ii), we have already produced results in simpler cases (without taking into account failures), and we also have results on all the associated areas (dependability analysis, combinatorial optimization, etc.). These main research lines are completed with other goals all concerned with the quantitative analysis of such complex communication systems.

• Program: Math AmSud

• Title: Stochastic Analysis, Statistics Inference, Numerical Analysiss (SIN)

• Inria principal investigator: Gerardo Rubino

• Main International Partners (Institution - Laboratory - Researcher):

  • University of the Republic (UDELAR), Montevideo, Uruguay – Computer Science at the Engineering Faculty – prof. Paola Bermolen

  • University of Valparaíso, Chile — Prof. Soledad Torres

• Duration: 2 years, Jan. 2013 – Dec. 2014

• Stochastic calculus with respect to the standard Brownian motion or more generally with respect to semi-martingales is currently one of the most important components of international research in probability theory. The applications of this theory largely exceed the original probabilistic framework and have repercussions in various fields, including differential geometry, differential partial equations, theoretical physics, modeling in finance, hydrology, telecommunications and biology. Recently, many authors have been interested in developing a stochastic calculus with respect to Gaussian processes which are not necessarily semi-martingales, as for instance the well known fractional Brownian motion. This research project is articulated around the analysis and the applications of stochastic differential equations driven by long memory processes.

  SIN is a large project with many partners. Our team participates in looking at differential equations and stocastic differential equations as limits of discrete Markov processes.