This section lists the available contributions for Opus. When applicable, they can be downloaded from the 'download' link below.

Opus-Lib:

  • Feel++ for Opus, v.0.91 and 0.92.0 : Feel++-OPUS is a framework for the Certified Reduced Basis (CRB) methods presented during the first OPUS workshop. It essentially provides a C++ interface for finite element codes which only supports Feel++ and an implementation of some CRB methods. See here for a pdf presentation.

  • RPyWrapper : the wrapper RPyWrap is a link module which allows to make software contributions developed in language R interoperable with Open TURNS.
  • Open TURNS PC library : Open TURNS is a C++ library and a Python module that implements the global methodology for uncertainty propagation that Opus also uses. The Opus project created new features for Openturns, especially a Polynomial Chaos Expansion (PCE) library. This is now used on real industrial cases on a regular basis and is available for the whole community for business, academic or teaching purposes. See website and blog for details. This is a joint work of EADS, EDF and Phiméca. The contribution of EADS to these developments has been mainly funded by OPUS, while the contribution of EDF and Phiméca has been funded by own resources and by the ANR project MIRADOR (Modélisation interactive des risques associés au développement d’ouvrages robustes, ref. ANR-06-RGCU-008).
  • NISP toolbox for Scilab : The goal of this toolbox is to provide a tool to manage uncertainties in simulated models for the Scilab platform. This Scilab toolbox is based on the NISP C++ library, where NISP stands for "Non-Intrusive Spectral Projection". The NISP library is based on a set of 3 C++ classes so that it provides an object-oriented framework for uncertainty analysis. The Scilab toolbox provides a pseudo-object oriented interface to this library, so that the two approaches are consistent. See wiki for details.

Opus Contrib :

  • Octave/Matlab toolbox for kriging : a toolbox for carrying out kriging regression in Matlab (also compatible with Octave, with some limitations). See website on sourceforge for more details.
  • Various R scripts for inverse probabilistic modeling were created, dealing with :
    • MCMC estimation usecase scripts: The idea here was to develop the scripts applied to a usecase while reusing existing tools based on the R language when relevant. Two R packages available on the CRAN site have been identified, and both packages may be used conjointly with the R package CODA which allows output analysis and diagnostics for Markov Chain Monte Carlo simulations.These packages are MCMC package and MCMCpack package.
    • MLE estimation scripts:
      • ECME (Expectation Conditional Maximization Either) by iterated linearizations (to deal with non-linear cases), based on G. Celeux et al. (Identifying intrinsic variability in multivariate systems through linearised inverse methods, Rapport de recherche INRIA RR-6400, 2007);
      • S(A)EM (Stochastic - Approximation version of - Expectation Maximization) which allows to perform SEM or SAEM.
      • Relevant publications to SA(E)M : G. Celeux and J. Diebolt, The SEM Algorithm: a Probabilistic 101/128 ANR OPUS Final Report D-WP0/11/03/A Teacher Algorithm Derived from the EM Algorithm for the Mixture Problem, Computational Statistics Quaterly, 2:73–82, 1985 and E. Kuhn, Estimation par maximum de vraisemblance dans des problèmes inverses non linéaires, Mémoire de thèse, Université Paris XI d’Orsay, 2003).
Opus-Forum :
  • High quantile estimation by Multi-Element Polynomial Chaos expansions. This contribution comes mainly from the work of the post-doc Jordan Ko (Paris Diderot). It is a brief summary of a paper submitted to Journal of Computational Physics.
  • Functional sensitivity analysis: variable selection using varying coefficient modeling and application to diesel engine smoke depollution. This contribution came mainly from the works of the DICE consortium. It is thus a contribution to the platform, ’external’ to the OPUS project team. A commented R script implementing the proposed method, as well as data files have been provided.

Download contributions

< Prev   Next >

Last Updated (Monday, 26 September 2011 02:15)