Search

International audience; We consider the damped nonlinear wave (NLW) equation driven by a noise which is white in time and colored in space. Assuming that the noise is non-degenerate in all Fourier modes, we establish a large deviations principle (LDP) for the occupation measures of the trajectories. The lower bound in the LDP is of a local type, which is related to the weakly dissipative nature of the equation and is a novelty in the context of randomly forced PDE’s. The proof is based on an extension of methods developed in (Comm. Pure Appl. Math. 68 (12) (2015) 2108–2143) and (Large deviations and mixing for dissipative PDE’s with unbounded random kicks (2014) Preprint) in the case of kick forced dissipative PDE’s with parabolic regularization property such as, for example, the Navier–Stokes system and the complex Ginzburg–Landau equations. We also show that a high concentration towards the stationary measure is impossible, by proving that the rate function that governs the LDP cannot have the trivial form (i.e., vanish on the stationary measure and be infinite elsewhere).

Density dependent Markov population processes in large populations of size $N$ were shown by Kurtz (1970, 1971) to be well approximated over finite time intervals by the solution of the differential equations that describe their average drift, and to exhibit stochastic fluctuations about this deterministic solution on the scale $\sqrt N$ that can be approximated by a diffusion process. Here, motivated by an example from evolutionary biology, we are concerned with describing how such a process leaves an absorbing boundary. Initially, one or more of the populations is of size much smaller than $N$, and the length of time taken until all populations have sizes comparable to $N$ then becomes infinite as $N \to \infty$. Under suitable assumptions, we show that in the early stages of development, up to the time when all populations have sizes at least $N^{1-\alpha}$, for $1/3 < \alpha < 1$, the process can be accurately approximated in total variation by a Markov branching process. Thereafter, the process is well approximated by the deterministic solution starting from the original initial point, but with a random time delay. Analogous behaviour is also established for a Markov process approaching an equilibrium on a boundary, where one or more of the populations become extinct., Comment: 50 pages

International audience; In the continuity of a recent paper ([6]), dealing with finite Markov chains, this paper proposes and analyzes a recursive algorithm for the approximation of the quasi-stationary distribution of a general Markov chain living on a compact metric space killed in finite time. The idea is to run the process until extinction and then to bring it back to life at a position randomly chosen according to the (possibly weighted) empirical occupation measure of its past positions. General conditions are given ensuring the convergence of this measure to the quasi-stationary distribution of the chain. We then apply this method to the numerical approximation of the quasi-stationary distribution of a diffusion process killed on the boundary of a compact set and to the estimation of the spectral gap of irreducible Markov processes. Finally, the sharpness of the assumptions is illustrated through the study of the algorithm in a non-irreducible setting.

In this paper, we investigate the fluctuations of a unit eigenvector associated to an outlier in the spectrum of a spiked $N\times N$ complex Deformed Wigner matrix $M_N$: $M_N =W_N/\sqrt{N} + A_N$ where $W_N$ is an $N \times N$ Hermitian Wigner matrix whose entries have a law $\mu$ satisfying a Poincar\'e inequality and the matrix $A_N$ is a block diagonal matrix, with an eigenvalue $\theta$ of multiplicity one, generating an outlier in the spectrum of $M_N$. We prove that the fluctuations of the norm of the projection of a unit eigenvector corresponding to the outlier of $M_N$ onto a unit eigenvector corresponding to $\theta$ are not universal., Comment: accepted for publication in Latin American Journal of Probability and Mathematical Statistics

International audience; We propose in this paper a construction of a diffusion process on the Wasserstein space P_2(R) of probability measures with a second-order moment. This process was introduced in several papers by Konarovskyi (see e.g. "A system of coalescing heavy diffusion particles on the real line", 2017) and consists of the limit when N tends to infinity of a system of N coalescing and mass-carrying particles. It has properties analogous to those of a standard Euclidean Brownian motion, in a sense that we will precise in this paper. We also compare it to the Wasserstein diffusion on P_2(R) constructed by von Renesse and Sturm (see Entropic measure and Wasserstein diffusion). We obtain that process by the construction of a system of particles having short-range interactions and by letting the range of interactions tend to zero. This construction can be seen as an approximation of the singular process of Konarovskyi by a sequence of smoother processes.

We derive distributional limits for empirical transport distances between probability measures supported on countable sets. Our approach is based on sensitivity analysis of optimal values of infinite dimensional mathematical programs and a delta method for non-linear derivatives. A careful calibration of the norm on the space of probability measures is needed in order to combine differentiability and weak convergence of the underlying empirical process. Based on this we provide a sufficient and necessary condition for the underlying distribution on the countable metric space for such a distributional limit to hold. We give an explicit form of the limiting distribution for ultra-metric spaces. Finally, we apply our findings to optimal transport based inference in large scale problems. An application to nanoscale microscopy is given.

We prove that, for (adapted) stationary processes, the so-called Maxwell–Woodroofe condition is sufficient for the law of the iterated logarithm and that it is optimal in some sense. That result actually holds in the context of Banach valued stationary processes, including the case of $L^{p}$-valued random variables, with $1\le p

The Ewens sampling formula was firstly introduced in the context of population genetics by Warren John Ewens in 1972, and has appeared in a lot of other scientific fields. There are abundant approximation results associated with the Ewens sampling formula especially when one of the parameters, the sample size $n$ or the mutation parameter $\theta$ which denotes the scaled mutation rate, tends to infinity while the other is fixed. By contrast, the case that $\theta$ grows with $n$ has been considered in a relatively small number of works, although this asymptotic setup is also natural. In this paper, when $\theta$ grows with $n$, we advance the study concerning the asymptotic properties of the total number of alleles and of the counts of components in the allelic partition assuming the Ewens sampling formula from the viewpoint of Poisson approximations., Comment: 38 pages

We consider random fields indexed by finite subsets of an amenable discrete group, taking values in the Banach-space of bounded right-continuous functions. The field is assumed to be equivariant, local, coordinate-wise monotone, and almost additive, with finite range dependence. Using the theory of quasi-tilings we prove an uniform ergodic theorem, more precisely, that averages along a Foelner sequence converge uniformly to a limiting function. Moreover we give explicit error estimates for the approximation in the sup norm., Comment: 28 pages, 1 figure, accepted in Annales of Applied Probability

We combine the method of exchangeable pairs with Stein's method for functional approximation. As a result, we give a general linearity condition under which an abstract Gaussian approximation theorem for stochastic processes holds. We apply this approach to estimate the distance of a sum of random variables, chosen from an array according to a random permutation, from a Gaussian mixture process. This result lets us prove a functional combinatorial central limit theorem. We also consider a graph-valued process and bound the speed of convergence of the distribution of its rescaled edge counts to a continuous Gaussian process., Comment: will appear in Annales de l'Institut Henri Poincar\'e, Probabilit\'es et Statistiques

We study the properties of the multivariate skew normal distribution as an approximation to the distribution of the sum of n independent, identically distributed random vectors. More precisely, we establish conditions ensuring that the uniform distance between the two distribution functions converges to 0 at a rate of n -2/3. The advantage over the corresponding normal approximation is particularly relevant when the summands are skewed and n is small, as illustrated for the special case of exponentially distributed random variables. Applications to some well-known multivariate distributions are also discussed.

We introduce the multiplicative coalescent with linear deletion, a continuous-time Markov process describing the evolution of a collection of blocks. Any two blocks of sizes $x$ and $y$ merge at rate $xy$, and any block of size $x$ is deleted with rate $\lambda x$ (where $\lambda\geq 0$ is a fixed parameter). This process arises for example in connection with a variety of random-graph models which exhibit self-organised criticality. We focus on results describing states of the process in terms of collections of excursion lengths of random functions. For the case $\lambda=0$ (the coalescent without deletion) we revisit and generalise previous works by authors including Aldous, Limic, Armendariz, Uribe Bravo, and Broutin and Marckert, in which the coalescence is related to a "tilt" of a random function, which increases with time; for $\lambda>0$ we find a novel representation in which this tilt is complemented by a "shift" mechanism which produces the deletion of blocks. We describe and illustrate other representations which, like the tilt-and-shift representation, are "rigid", in the sense that the coalescent process is constructed as a projection of some process which has all of its randomness in its initial state. We explain some applications of these constructions to models including mean-field forest-fire and frozen-percolation processes., Comment: 45 pages, 6 figures. Updated references

In this article, we study weighted particle representations for a class of stochastic partial differential equations (SPDE) with Dirichlet boundary conditions. The locations and weights of the particles satisfy an infinite system of stochastic differential equations. The locations are given by independent, stationary reflecting diffusions in a bounded domain, and the weights evolve according to an infinite system of stochastic differential equations driven by a common Gaussian white noise $W$ which is the stochastic input for the SPDE. The weights interact through $V$, the associated weighted empirical measure, which gives the solution of the SPDE. When a particle hits the boundary its weight jumps to a value given by a function of the location of the particle on the boundary. This function determines the boundary condition for the SPDE. We show existence and uniqueness of a solution of the infinite system of stochastic differential equations giving the locations and weights of the particles and derive two weak forms for the corresponding SPDE depending on the choice of test functions. The weighted empirical measure $V$ is the unique solution for each of the nonlinear stochastic partial differential equations. The work is motivated by and applied to the stochastic Allen-Cahn equation and extends the earlier of work of Kurtz and Xiong in [14, 15].

In a recent paper Velleman and Warrington analyzed the expected values of some of the parameters in a memory game, namely, the length of the game, the waiting time for the first match, and the number of lucky moves. In this paper we continue this direction of investigation and obtain the limiting distributions of those parameters. More specifically, we prove that when suitably normalized, these quantities converge in distribution to a normal, Rayleigh, and Poisson random variable, respectively. We also make a connection between the memory game and one of the models of preferential attachment graphs. In particular, as a by--product of our methods we obtain simpler proofs (although without rate of convergence) of some of the results of Pek\"oz, R\"ollin, and Ross on the joint limiting distributions of the degrees of the first few vertices in preferential attachment graphs. For proving that the length of the game is asymptotically normal, our main technical tool is a limit result for the joint distribution of the number of balls in a multi--type generalized P\'olya urn model., Comment: 24 pages, 1 figure

The instant system availability \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$S_\tau (t)$$\end{document}Sτ(t) of a repairable system with the renewal equation was studied. The starting point monotonicity of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$S_\tau (t)$$\end{document}Sτ(t) was proved and the upper bound of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$S_\tau (t)$$\end{document}Sτ(t) is also derived. It was found that the interval of instant system availability monotonically decreases. In addition, we provide examples that validate the analytically derived properties of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$S_\tau (t)$$\end{document}Sτ(t) based on the Lognormal, Gamma and Weibull distributions and the results show that the value of T is slightly smaller than its value defined in Theorem 2. The procedure of using a bathtub as application for this article is also discussed.

We present a class of function spaces valued pth-order autoregressive random processes and we study their main properties. We also give examples of special functional autoregressive representation (with not strong white noise) which is different of that given in Bosq [6]. This functional approach seems a promising tool to investigate statistical applications of continuous time random process and to obtain new limit theorems for dependent variables.

We study the limiting probability distribution of the homogenization error for second order elliptic equations in divergence form with highly oscillatory periodic conductivity coefficients and highly oscillatory stochastic potential. The effective conductivity coefficients are the same as those of the standard periodic homogenization, and the effective potential is given by the mean. We show that in the short range correlation setting, the limiting distribution of the random part of the homogenization error, as random elements in proper Hilbert spaces, is Gaussian and can be characterized by the homogenized Green's function, the homogenized solution and the statistics of the random potential. Similar results hold for random potentials that are functions of long range correlated Gaussian random fields. These results generalize previous ones in the setting with slowly varying diffusion coefficients, and the current setting with fast oscillations in the differential operator requires new methods to prove compactness of the probability distributions of the random fluctuation., 41 pages, 2 figures

This thesis is devoted to limit theorems for strictly stationary sequences and random fields. We concentrate essentially on the central limit theorem and its invariance principle.First, we show with the help of a counter-example that for a strictly stationary absolutely regular sequence, the central limit theorem may hold but not the invariance principle. We also show that the central limit theorem does not take place for partial sums of a Hilbert space valued, strictly stationary and absolutely regular sequence, even if we assume that the normalized partial sums form a uniformly integrable family. Second, we investigate the Holderian invariance principle. We treat the case of $\tau$-dependent (Dedecker, Prieur, 2005) and $\rho$-mixing strictly stationary sequences. We provide a sufficient condition on the law of a strictly stationary martingale difference sequence and the quadratic variance which guarantee the invariance principle in a Hölder space. We construct a counter-example which shows its sharpness. We derive conditions in the spirit of Hannan (1979), and Maxwell and Woodroofe (2000) by a martingale approximation.We then discuss the martingale/coboundary decomposition. In dimension one, we provide sharp integrability conditions on the transfer function and the coboundary for which the later does not spoil the invariance principle, the law of the iterated logarithm or the strong law of large numbers if these theorems take place for the martingale involved in the decomposition. We also provide a sufficient condition for an orthomartingale/coboundary decomposition for strictly stationary random fields.Lastly, we establish tails inequalities for orthomartingale and Bernoulli random fields. We prove an invariance principle in Hölder spaces for these random fields using such inequalities., Cette thèse est consacrée aux théorèmes limites pour les suites et les champs aléatoires strictement stationnaires. Nous étudions essentiellement le théorème limite central et sa version fonctionnelle, appelée principe d'invariance. Dans un premier temps, nous montrons à l'aide d'un contre-exemple que pour les processus strictement stationnaires $\beta$-mélangeants, le théorème limite central peut avoir lieu sans que ce ne soit le cas pour la version fonctionnelle. Nous montrons également que le théorème limite central n'a pas nécessairement lieu pour les sommes partielles d'une suite strictement stationnaire $\beta$-mélangeante à valeurs dans un espace de Hilbert de dimension infinie, même en supposant l'uniforme intégrabilité de la suite des sommes partielles normalisées.Puis nous étudions le principe d'invariance dans l'espace des fonctions hölderiennes. Nous traitons le cas des suites strictement stationnaires $\tau$-dépendantes (au sens de Dedecker, Prieur, 2005) ou $\rho$-mélangeantes. Nous donnons également une condition suffisante sur la loi d'une suite strictement stationnaire d'accroissements d'une martingale et la variance conditionnelle garantissant le principe d'invariance dans l'espace des fonctions hölderiennes, et nous démontrons son optimalité à l'aide d'un contre-exemple. Ensuite, nous déduisons grâce à une approximation par martingales des conditions dans l'esprit de celles de Hannan (1979), et Maxwell et Woodroofe (2000).Nous discutons ensuite de la décomposition martingale/cobord. Dans le cas des suites, nous fournissons des conditions d'intégrabilité sur la fonction de transfert et le cobord pour que ce dernier ne perturbe pas le principe d'invariance, la loi des logarithmes itérés ou bien la loi forte des grands nombres si ceux-ci ont lieu pour la martingale issue de la décomposition. Dans le cas des champs, nous formulons une condition suffisante pour une décomposition ortho-martingale/cobord. Enfin, nous établissons des inégalités sur les queues des maxima des sommes partielles d'un champ aléatoire de type ortho-martingale ou bien d'un champ qui s'exprime comme une fonctionnelle d'un champ i.i.d. Ces inégalités permettent d'obtenir un principe d'invariance dans les espaceshölderiens pour ces champs aléatoires.

Dans (Ann. Inst. Henri Poincare Probab. Stat. 34 (1998) 637–686), les auteurs obtiennent une chaine de Markov a valeurs arbres en elaguant de plus en plus de sous-arbres le long des nœuds d’un arbre de Galton–Watson. Plus recemment dans (Ann. Probab. 40 (2012) 1167–1211), un analogue continu de la dynamique d’elagage a valeurs arbres est construit sur des arbres de Levy. Dans cet article, nous presentons une nouvelle topologie qui permet de relier les dynamiques discretes et continues en les considerant comme des exemples du meme processus de Markov fort avec des conditions initiales differentes. Nous construisons ce processus d’elagage sur l’espace des arbres appeles bi-mesures, qui sont des espaces metriques mesures avec une mesure d’elagage additionnelle. La mesure d’elagage est supposee finie sur les arbres finis, mais pas necessairement localement finie. De plus, nous caracterisons analytiquement le processus d’elagage par son generateur infinitesimal et montrons qu’il est continu en son arbre bi-mesure initial. Plusieurs exemples sont donnes, notamment le cas d’une loi de reproduction a variance finie ou la mesure d’elagage est la mesure des longueurs sur l’arbre sous-jacent.

In this paper, we consider sequences of vector martingale differences of increasing dimension. We show that the Kantorovich distance from the distribution of the k(n)-dimensional average of n martingale differences to the corresponding Gaussian distribution satisfies certain inequalities. As a consequence, if the growth of k(n) is not too fast, then the Kantorovich distance converges to zero. Two applications of this result are presented. The first is a precise proof of the asymptotic distribution of the multivariate portmanteau statistic applied to the residuals of an autoregressive model and the second is a proof of the asymptotic normality of the estimates of a finite autoregressive model when the process is an AR(∞) and the order of the model grows with the length of the series.

Permanent link to this document: http://projecteuclid.org/euclid.ojm/1216151109; International audience; In a recent paper, Nicolas Bouleau provides a new tool, based on the language of Dirichlet forms, to study the propagation of errors and reinforce the historical approach of Gauss. In the same way that the practical use of the normal distribution in statistics may be explained by the central limit theorem, the aim of this paper is to underline the importance of a family of error structures by asymptotic arguments.

In this paper, we establish asymptotic expansions for the Laplace approximations for Ito functionals of Brownian rough paths under the condition that the phase function has finitely many non-degenerate minima. Our main tool is the Banach space-valued rough path theory of T. Lyons. We use a large deviation principle and the stochastic Taylor expansion with respect to the topology of the space of geometric rough paths. This is a continuation of a series of papers by Inahama [Y. Inahama, Laplace's method for the laws of heat processes on loop spaces, J. Funct. Anal. 232 (2006) 148–194] and by Inahama and Kawabi [Y. Inahama, H. Kawabi, Large deviations for heat kernel measures on loop spaces via rough paths, J. London Math. Soc. 73 (3) (2006) 797–816], [Y. Inahama, H. Kawabi, On asymptotics of certain Banach space-valued Ito functionals of Brownian rough paths, in: Proceedings of the Abel Symposium 2005, Stochastic Analysis and Applications, A Symposium in Honor of Kiyosi Ito, Springer, Berlin, in press. Available at: http://www.abelprisen.no/no/abelprisen/deltagere_2005.html ].

Swiss National Foundation : FN 200020-149871/1. CIMI (Centre International de Mathematiques et d'Informatique) : ANR-11-LABX-0040-CIMI, ANR-11-IDEX-0002-02; International audience; This work is concerned with the analysis of a stochastic approximation algorithm for the simulation of quasi-stationary distributions on finite state spaces. This is a generalization of a method introduced by Aldous, Flannery and Palacios. It is shown that the asymptotic behavior of the empirical occupation measure of this process is precisely related to the asymptotic behavior of some deterministic dynamical system induced by a vector field on the unit simplex. This approach provides new proof of convergence as well as precise asymptotic rates for this type of algorithm. In the last part, our convergence results are compared with those of a particle system algorithm (a discrete-time version of the Fleming-Viot algorithm).