BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Space filling designs for computer experiments: some algorithms an
 d numerical results on industrial problems - Iooss\, B (EDF R&amp\;D)
DTSTART:20110906T090000Z
DTEND:20110906T093000Z
UID:TALK32675@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:Complex computer codes\, for instance simulating physical phen
 omena\, are often too time expensive to be directly used to perform uncert
 ainty\, sensitivity\, optimization and robustness analyses. A widely accep
 ted method to circumvent this problem consists in replacing cpu time expen
 sive computer models by cpu inexpensive mathematical functions\, called me
 tamodels. A necessary condition to a successful modelling of these compute
 r experiments is to explore the whole space of variations of the computer 
 model input variables. However in many industrial applications\, we are fa
 ced to the harsh problem of the high dimensionality of the exploration spa
 ce. In this communication\, we will first focus on the metamodel validatio
 n process which consists in evaluating the metamodel predictivity with res
 pect to the initial computer code. This step has to be realized with cauti
 on and robustness in industrial applications\, especially in the framework
  of safety studies. \n\nWe propose and test an algorithm\, which optimizes
  the distance between the validation points and the metamodel learning poi
 nts in order to estimate the true metamodel predictivity with a minimum nu
 mber of validation points. Comparisons with classical validation algorithm
 s and application to a nuclear safety computer code show the relevance of 
 this sequential validation design. Second\, we will present some recent re
 sults about the properties of different space filling designs. In practice
 \, one has to choose which design to use in an exploratory phase of a nume
 rical model. We will show the usefulness of some classification tools\, as
  those based on the minimal spanning trees. We adopt a numerical approach 
 to compare the performance of different types of space filling designs\, i
 n function of their interpoint-distance\, L2-discrepancies and various sub
 -projection properties. Finally\, we will present two recent problems\, po
 sed in some industrial applications: the introductions of inequality const
 raints between the inputs of a space filling design and the building of sp
 ace filling design mixing quantitative and qualitative factors. \n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR