Concentration Inequalities for Mean Field Particle Models
نویسندگان
چکیده
This article is concerned with the fluctuations and the concentration properties of a general class of discrete generation and mean field particle interpretations of non linear measure valued processes. We combine an original stochastic perturbation analysis with a concentration analysis for triangular arrays of conditionally independent random sequences, which may be of independent interest. Under some additional stability properties of the limiting measure valued processes, uniform concentration properties with respect to the time parameter are also derived. The concentration inequalities presented here generalize the classical Hoeffding, Bernstein and Bennett inequalities for independent random sequences to interacting particle systems, yielding very new results for this class of models. We illustrate these results in the context of McKean Vlasov type diffusion models, McKean collision type models of gases, and of a class of Feynman-Kac distribution flows arising in stochastic engineering sciences and in molecular chemistry. Key-words: Concentration inequalities, mean field particle models, measure valued processes, Feynman-Kac semigroups, McKean Vlasov models. ∗ Centre INRIA Bordeaux et Sud-Ouest & Institut de Mathématiques de Bordeaux , Université de Bordeaux I, 351 cours de la Libération 33405 Talence cedex, France, [email protected] † Centre INRIA Bordeaux et Sud-Ouest & LMV Université de Versailles Bâtiment Fermat, 45 Av. des Etats-Unis, 78035 Versailles Cedex France, [email protected] in ria -0 03 75 13 4, v er si on 3 26 A pr 2 00 9 Inégalités de concentration pour des modèles particulaires de champ moyen Résumé : Nous analysons dans cet article les fluctuations et les propriétés de concentration d’une classe générale de systèmes de particules en interaction de type champ moyen et à temps discret. Ces modèles probabilistes sont liés à des interprétations particulaires de processus à valeurs mesures non linéaires. Nous développons une analyse originale fondée sur des techniques de perturbation stochastique de semigroupes non linéaires et sur un théorème de fluctuations de tableaux triangulaires de variables conditionnellement indépendantes. Dans certaines conditions de stabilité des semigroupes associés au processus limite, nous présentons des inégalités de concentration uniformes par rapport au paramètre temporel. Les inégalités de concentration développées dans cette étude sont des extensions des inégalités classiques de Hoeffding, Bernstein et de Bennett dans le cadre des sequence de variables indépendantes, à des systèmes de particules en interaction. Ces résultats semblent être les premiers de ce type pour ces classes de processus en interaction. Nous illustrons ces propriétés de concentration dans le cadre de modèles diffusifs de type McKean Vlasov, pour des modèles de collisions de type McKean issus de la mécanique des fluides, ainsi que pour une classe de modèles de Feynman-Kac utilisés en ingénierie stochastique et en chimie moléculaire. Mots-clés : Inégalités de concentration, modèles particulaires de champ moyen, processus à valeurs mesures, semigroupes de Feynman-Kac, modèles de McKean Vlasov in ria -0 03 75 13 4, v er si on 3 26 A pr 2 00 9 Concentration Inequalities for Mean Field Particle Models 3
منابع مشابه
Sino-French Summer Institute: Stochastic Modeling and Applications
of Mini-course On the concentration properties of mean field particle models (based on a series of 4 joint works with : D.A. Dawson, A. Guionnet, E. Rio, S.L. Hu and L.M. Wu) Pierre Del Moral Centre de Recherche Bordeaux Sud-Ouest & Institut de Mathématique de Bordeaux INRIA, France Abstract This lecture is concerned with the exponential concentration properties of a general class of mean field...
متن کاملQuantitative Concentration Inequalities for Empirical Measures on Non-compact Spaces
We establish some quantitative concentration estimates for the empirical measure of many independent variables, in transportation distances. As an application, we provide some error bounds for particle simulations in a model mean field problem. The tools include coupling arguments, as well as regularity and moments estimates for solutions of certain diffusive partial differential equations.
متن کاملLogarithmic Sobolev inequalities for some nonlinear PDE’s
The aim of this paper is to study the behavior of solutions of some nonlinear partial differential equations of Mac Kean-Vlasov type. The main tools used are, on one hand, the logarithmic Sobolev inequality and its connections with the concentration of measure and the transportation inequality with quadratic cost; on the other hand, the propagation of chaos for particle systems in mean field in...
متن کاملOn the stability and the uniform propagation of chaos properties of Ensemble Kalman-Bucy filters
The Ensemble Kalman filter is a sophisticated and powerful data assimilation method for filtering high dimensional problems arising in fluid mechanics and geophysical sciences. This Monte Carlo method can be interpreted as a mean-field McKean-Vlasov type particle interpretation of the Kalman-Bucy diffusions. In contrast to more conventional particle filters and nonlinear Markov processes these ...
متن کاملEffect of Sweeteners on Viscosity and Particle Size of Dilute Guar Gum Solutions
The effects of some synthetic sweeteners on the rheological and physical properties of guar gum in dilute solutions were investigated.Measurements include the determination of intrinsic viscosity and the particle size, surface weighted mean [D3, 2], volume weighted mean [D4,3] and specificsurface area ofguar gum andsyntheticsweeteners mixtures. T...
متن کامل