The research topic is continuous stochastic optimization and is at the crossroad between between probability theory, numerical analysis and computer science. The principal goal is to develop novel methodologies to support simulation-based model building and design. In this context the task of the thesis is threefold. I. Derive linear convergence results for state-of-the-art adaptive stochastic search algorithms for classes of functions that include noise. Here, previous theoretical results exploiting the connection between MCMC methods and adaptive stochastic search algorithms will be used.
II. Exploit and integrate several statistical machine learning techniques in the context of stochastic optimization. The main objectives are to decrease the number of internal parameters and therefore the learning time and to introduce non-linear internal models. III. Improve and provide new developments of an environment for benchmarking stochastic optimization algorithms.
Context
Objectives
Work program
Extra information
Prerequisite
Détails
Expected funding
Institutional funding
Status of funding
Expected
Candidates
Alexandre Chotard
Utilisateur
Créé
Mardi 07 juin 2011 15:03:58 CEST
dernière modif.
Lundi 20 juin 2011 15:38:36 CEST
Fichiers joints
filename
créé
hits
filesize
Aucun fichier joint à cette fiche
Connexion
Ecole Doctorale Informatique Paris-Sud
Directrice
Nicole Bidoit Assistante
Stéphanie Druetta Conseiller aux thèses
Dominique Gouyou-Beauchamps
ED 427 - Université Paris-Sud
UFR Sciences Orsay
Bat 650 - aile nord - 417
Tel : 01 69 15 63 19
Fax : 01 69 15 63 87
courriel: ed-info à lri.fr