Theme Metamodeling
Presentation
The MOCA team studies the management of numerical simulation models and their use in continuous optimization. Metamodeling encompasses here the simulation of physical systems and the two types of metamodels (models of models) :
-
Statistical metamodels (e.g., Gaussian processes) build from input-output data. The data often comes itself from the simulation of physical systems (e.g., finite elements models). We specialize in statistical metamodels built from sparse data. Such metamodels typically underly optimization algorithms.
-
Metamodels that aim at producing computer programs by describing and processing a class of models. Such metamodels may also be metacodes that algorithmically describe a set of possible program instances.
The group has contributions in domains such as continuous optimization theory and computer experiments. In addition, the group implements simulation softwares for high-performance computing, discrete events simulation software, and focuses on the numerical reproductibility of the results.
Keywords:
- Statistical models
- Continuous optimization
- Discrete event simulation
- Scientific Computing
- High Performance Computing
- Numerical reproductibility
Last publications
Lagrangian bounds for large‐scale multicommodity network design: a comparison between Volume and Bundle methods
International Transactions in Operational Research
Jhouben Cuesta Ramirez, Olivier Roustant, Alain Gliere, Rodolphe Le Riche, Cédric Durantin, Guillaume Perrin - Dec. 17, 2020
Optimization in presence of categorical variables
Journées scientifiques d'automne de la chaire OQUAIDO
Julien Pelamatti, Rodolphe Le Riche, Celine Helbert, Christophette Blanchet-Scalliet - Dec. 17, 2020
Coupling constraints in Bayesian optimization with uncertainties
Journées scientifiques d'automne de la chaire OQUAIDO
Olivier Roustant, Rodolphe Le Riche - Dec. 17, 2020
de OQUAIDO à CIROQUO -- bilan de 5 ans de la chaire OQUAIDO en mathématiques appliquées
Journées scientifiques d'automne de la chaire OQUAIDO
Alexandre Bardakoff, Bruno Bachelet, Timothy Blattner, Walid Keyrouz, Gerson C. Kroiz, Loïc Yon - Nov. 12, 2020
Hedgehog: Understandable Scheduler-Free Heterogeneous Asynchronous Multithreaded Data-Flow Graphs
IEEE/ACM 3rd Annual Parallel Applications Workshop: Alternatives To MPI+X (PAW-ATM)
Alexis Pereda, David R.C. Hill, Claude Mazel, Bruno Bachelet - Oct. 21, 2020
Repeatability with Random Numbers Using Algorithmic Skeletons
34th European Simulation and Modelling Conference (ESM)
Nicolas Wagner, Violaine Antoine, Jonas Koko, Marie-Madeleine Mialon, Romain Lardy, Isabelle Veissier - Sept. 17, 2020
Comparison of Machine Learning Methods to Detect Anomalies in the Activity of Dairy Cows
Rodolphe Le Riche, Adrien Spagnol, David Gaudrie, Sébastien da Veiga, Victor Picheny - July 9, 2020
Reducing dimension in Bayesian Optimization
LIMOS internal seminar
Nicolas Wagner, Violaine Antoine, Jonas Koko, Romain Lardy - June 15, 2020
Fuzzy k-NN Based Classifiers for Time Series with Soft Labels
18. International Conference on Information Processing and Management of Uncertainty in Knowledge-Based Systems
Alexandre Bardakoff, Timothy Blattner, Bruno Bachelet, Walid Keyrouz, Loïc Yon - March 22, 2020
Hedgehog: a Performance-Oriented General Purpose Library for Multi-GPU Systems
GPU Technology Conference (GTC)
All publications are here