%0 Journal Article
%T An Optimization Strategy Based on the Maximization of Matching-Targets¡¯ Probability for Unevaluated Results Optimisation par maximisation de la probabilit¨¦ d¡¯atteindre les cibles pour des r¨¦sultats non ¨¦valu¨¦s
%A Feraille M.
%J Oil & Gas Science and Technology
%D 2013
%I Institut Fran?ais du P¨¦trole
%R 10.2516/ogst/2012079
%X The Maximization of Matching-Targets¡¯ Probability for Unevaluated Results (MMTPUR), technique presented in this paper, is based on the classical probabilistic optimization framework. The numerical function values that have not been evaluated are considered as stochastic functions. Thus, a Gaussian process uncertainty model is built for each required numerical function result (i.e., associated with each specified target) and is used to estimate probability density functions for unevaluated results. Parameter posterior distributions, used within the optimization process, then take into account these probabilities. This approach is particularly adapted when, getting one evaluation of the numerical function is very time consuming. In this paper, we provide a detailed outline of this technique. Finally, several test cases are developed to stress its potential. La m¨¦thode pr¨¦sent¨¦e dans cet article rentre dans le cadre de l¡¯optimisation probabiliste classique. La nouveaut¨¦ consiste ¨¤ construire un processus Gaussien pour chaque r¨¦sultat souhait¨¦ (i.e. associ¨¦ ¨¤ chaque cible sp¨¦cifi¨¦e) et de les utiliser ensuite pour estimer les densit¨¦s de probabilit¨¦ des r¨¦sultats non ¨¦valu¨¦s. Ces densit¨¦s sont alors prises en compte dans le calcul de la densit¨¦ a posteriori des param¨¨tres ¨¤ optimiser. Cette approche est adapt¨¦e lorsque chaque ¨¦valuation de la fonction est co teuse en temps de calcul. Une description d¨¦taill¨¦e de cette m¨¦thode d¡¯optimisation est propos¨¦e ainsi que son utilisation sur plusieurs cas tests.
%U http://dx.doi.org/10.2516/ogst/2012079