Let X = (X1,.....,Xm) be an infinitely degenerate system of vector fields, we study the existence and regularity of multiple solutions of Dirichelt problem for a class of semi-linear infinitely degenerate elliptic operators associated with the sum of square operator Δx = ∑m(j=1) Xj* Xj.

The topic of this thesis is a diffusion process on a potential landscape which is given by a smooth Hamiltonian function in the regime of small noise. The work provides a new proof of the Eyring-Kramers formula for the Poincaré inequality of the associated generator of the diffusion. The Poincaré inequality characterizes the spectral gap of the generator and establishes the exponential rate of convergence towards equilibrium in the L²-distance. This result was first obtained by Bovier et. al. in 2004 relying on potential theory. The presented approach in the thesis generalizes to obtain also asymptotic sharp estimates of the constant in the logarithmic Sobolev inequality. The optimal constant in the logarithmic Sobolev inequality characterizes the convergence rate to equilibrium with respect to the relative entropy, which is a stronger distance as the L²-distance and slightly weaker than the L¹-distance. The optimal constant has here no direct spectral representation. The proof makes use of the scale separation present in the dynamics. The Eyring-Kramers formula follows as a simple corollary from the two main results of the work: The first one shows that the associated Gibbs measure restricted to a basin of attraction has a good Poincaré and logarithmic Sobolev constants providing the fast convergence of the diffusion to metastable states. The second main ingredient is a mean-difference estimate. Here a weighted transportation distance is used. It contains the main contribution to the Poincaré and logarithmic Sobolev constant, resulting from exponential long waiting times of jumps between metastable states of the diffusion.

Ma thèse de doctorat se concentre principalement sur le comportement en temps long des processus de Markov, les inégalités fonctionnelles et les techniques relatives. Plus spécifiquement, Je vais présenter les taux de convergence sous-exponentielle explicites des processus de Markov dans deux approches : la méthode Meyn-Tweedie et l’hypocoercivité (faible). Le document se divise en trois parties. Dans la première partie, Je vais présenter quelques résultats importants et des connaissances connexes. D’abord, un aperçu de mon domaine de recherche sera donné. La convergence exponentielle (ou sous-exponentielle) des chaînes de Markov et des processus de Markov (à temps continu) est un sujet d’actualité dans la théorie des probabilité. La méthode traditionnelle développée et popularisée par Meyn-Tweedie est largement utilisée pour ce problème. Dans la plupart des résultats, le taux de convergence n’est pas explicite, et certains d’entre eux seront brièvement présentés. De plus, la fonction de Lyapunov est cruciale dans l’approche Meyn-Tweedie, et elle est aussi liée à certaines inégalités fonctionnelles (par exemple, inégalité de Poincaré). Cette relation entre fonction de Lyapounov et inégalités fonctionnelles sera donnée avec les résultats au sens L2. En outre, pour l’exemple de l’équation cinétique de Fokker-Planck, un résultat de convergence exponentielle explicite de la solution sera introduite à la manière de Villani : l’hypocoercivité. Ces contenus sont les fondements de mon travail, et mon but est d’étudier la décroissance sous-exponentielle. La deuxième partie, fait l’objet d’un article écrit en coopération avec d’autres sur les taux de convergence sous-exponentielle explicites des processus de Markov à temps continu. Comme nous le savons, les résultats sur les taux de convergence explicites ont été donnés pour le cas exponentiel. Nous les étendons au cas sous-exponentielle par l’approche Meyn-Tweedie. La clé de la preuve est l’estimation du temps de passage dans un ensemble ”petite”, obtenue par Douc, Fort et Guillin, mais pour laquelle nous donnons une preuve plus simple. Nous utilisons aussi la construction du couplage et donnons une ergodicité sous exponentielle explicite. Enfin, nous donnons quelques applications numériques. Dans la dernière partie, mon second article traite de l’équation cinétique de Fokker-Planck. Je prolonge l’hypocoercivité à l’hypocoercivité faible qui correspond à inégalité de Poincaré faible. Grâce à cette extension, on peut obtenir le taux de convergence explicite de la solution, dans des cas sous-exponentiels. La convergence est au sens H1 et au sens L2. A la fin de ce document, j’étudie le cas de l’entropie relative comme Villani, et j’obtiens la convergence au sens de l’entropie. Enfin, Je donne deux exemples pour les potentiels qui impliquent l’inégalité de Poincaré faible ou l’inégalité de Sobolev logarithmique faible pour la mesure invariante. / My Ph.D dissertation mainly focuses on long time behavior of Markov processes, functional inequalities and related techniques. More specifically, I will present the computable sub-exponential convergence rate of the Markov process in two approaches : Meyn-Tweedie’s method and (weak) hypocoercivity. The paper consists of three parts. In the first part, I will introduce some important results and related knowledge. Firstly, overviews of my research field are given. Exponential (or subexponential) convergence of Markov chains and (continuous time) Markov processes is a hot issue in probability. The traditional method - Meyn-Tweedie’s approach is widely applied for this problem. Most of the results about convergence rate is not explicit, and some of them will be introduced briefly. In addition,Lyapunov function is crucial in Meyn-Tweendie’s aproach, and it is also related to some functional inequalities (for example, Poincar´e inequality). The relationship of them will be given with results in L2 sense. Furthermore, as a example of kinetic Fokker-Planck equation, a computable result of exponential convergence of the solution of it will be introduced in Villani’ way - hypocoercivity. These contents are foundations of my work, and my destination is to study the sub-exponential decay. In the second part, it is my article cooperated with others about subexponential convergence rate of continuous time Markov processes. As we all know, the explicit results of convergence rate is about the exponential case. We extend them to sub-exponential case in Meyn-Tweedie’s approach. The key of the proof is the estimation of the hitting time to small set which was got by Douc, Fort and Guillin, for which we also propose an alternative simpler proof. We also use coupling construction as others and give a quantitative sub-exponential ergodicity. At last, we give some calculations for examples. In the last part, my second article deal with the kinetic Fokker-Planck equation. I extend the hypocoercivity to weak hypocoercivity which correspond to weak Poincar´e inequality. Through the extension, one can get the computable rate of convergence of the solution, which is also sub-exponential case. The convergence is in H1 sense and in L2 sense. In the end of this paper, I study the relative entropy case as C.Villani, and get convergence in entropy. Finally, I give two examples for potentials that implies weak Poincar´e inequality or weak logarithmic Sobolve inequality for invarient measure.

The topic of this thesis is a diffusion process on a potential landscape which is given by a smooth Hamiltonian function in the regime of small noise. The work provides a new proof of the Eyring-Kramers formula for the Poincaré inequality of the associated generator of the diffusion. The Poincaré inequality characterizes the spectral gap of the generator and establishes the exponential rate of convergence towards equilibrium in the L²-distance. This result was first obtained by Bovier et. al. in 2004 relying on potential theory. The presented approach in the thesis generalizes to obtain also asymptotic sharp estimates of the constant in the logarithmic Sobolev inequality. The optimal constant in the logarithmic Sobolev inequality characterizes the convergence rate to equilibrium with respect to the relative entropy, which is a stronger distance as the L²-distance and slightly weaker than the L¹-distance. The optimal constant has here no direct spectral representation. The proof makes use of the scale separation present in the dynamics. The Eyring-Kramers formula follows as a simple corollary from the two main results of the work: The first one shows that the associated Gibbs measure restricted to a basin of attraction has a good Poincaré and logarithmic Sobolev constants providing the fast convergence of the diffusion to metastable states. The second main ingredient is a mean-difference estimate. Here a weighted transportation distance is used. It contains the main contribution to the Poincaré and logarithmic Sobolev constant, resulting from exponential long waiting times of jumps between metastable states of the diffusion.