Cette thèse présente deux sujets de recherche indépendants concernant l'application de méthodes stochastiques à des problèmes issus de la dynamique moléculaire. Dans la première partie, nous présentons des travaux liés à l'interprétation probabiliste de l'équation de Poisson-Boltzmann qui intervient dans la description du potentiel électrostatique d'un système moléculaire. Après avoir introduit l'équation de Poisson-Boltzmann et les principaux outils mathématiques utilisés, nous nous intéressons à l'équation linéaire parabolique de Poisson-Boltzmann. Avant d’énoncer le résultat principal de la thèse, nous étendons des résultats d'existence et unicité des équations différentielles stochastiques rétrogrades. Nous donnons ensuite une interprétation probabiliste de l'équation non-linéaire de Poisson-Boltzmann sous la forme de la solution d'une équation différentielle stochastique rétrograde. Enfin, dans une seconde partie prospective, nous commençons l'étude d'une méthode proposée par Paul Malliavin de détection des variables lentes et rapides d'une dynamique moléculaire. / This thesis presents two independent research topics. Both are related to the application of stochastic problems to molecular dynamics. In the first part, we present a work related to the probabilistic interpretation of the Poisson-Boltzmann equation. This equation describes the electrostatic potential of a molecular system. After an introduction to the Poisson-Boltzmann equation, we focus on the parabolic and linear equation. After extending an existence and uniqueness result for backward stochastic differential equations, we establish a probabilistic interpretation of the nonlinear Poisson-Boltzmann equation with backward stochastic differential equations. Finally, in a more prospective second part, we initiate a study of a slow and fast variables detection method due to Paul Malliavin.

We partially solve the adaptive ergodic stochastic optimal control problem where the driving process is a fractional Brownian motion with Hurst parameter H > 1/2. A formula is provided for an optimal feedback control given a strongly consistent estimator of the parameters of the controlled system is avail- able. There are some special conditions imposed on the estimator which means the results are not completely general. They apply, for example, in the case where the estimator is independent of the driving fractional Brownian motion. In the course of the thesis, construction of stochastic integrals of suitable determinis- tic functions with respect to fractional Brownian motion with Hurst parameter H > 1/2 over the unbounded positive real half-line is presented as well. 1

Durant les dernières décennies, l'essor des moyens technologiques et particulièrement informatiques a permis l'émergence de la mise en œuvre de méthodes numériques pour l'approximation d'Equations Différentielles Stochastiques (EDS) ainsi que pour l'estimation de leurs paramètres. Cette thèse aborde ces deux aspects et s'intéresse plus spécifiquement à l'efficacité de ces méthodes. La première partie sera consacrée à l'approximation d'EDS par schéma numérique tandis que la deuxième partie traite l'estimation de paramètres. Dans un premier temps, nous étudions des schémas d'approximation pour les EDSs. On suppose que ces schémas sont définis sur une grille de temps de taille $n$. On dira que le schéma $X^n$ converge faiblement vers la diffusion $X$ avec ordre $h in mathbb{N}$ si pour tout $T>0$, $vert mathbb{E}[f(X_T)-f(X_T^n)] vertleqslant C_f /n^h$. Jusqu'à maintenant, sauf dans certains cas particulier (schémas d'Euler et de Ninomiya Victoir), les recherches sur le sujet imposent que $C_f$ dépende de la norme infini de $f$ mais aussi de ses dérivées. En d'autres termes $C_f =C sum_{vert alpha vert leqslant q} Vert partial_{alpha} f Vert_{ infty}$. Notre objectif est de montrer que si le schéma converge faiblement avec ordre $h$ pour un tel $C_f$, alors, sous des hypothèses de non dégénérescence et de régularité des coefficients, on peut obtenir le même résultat avec $C_f=C Vert f Vert_{infty}$. Ainsi, on prouve qu'il est possible d'estimer $mathbb{E}[f(X_T)]$ pour $f$ mesurable et bornée. On dit alors que le schéma converge en variation totale vers la diffusion avec ordre $h$. On prouve aussi qu'il est possible d'approximer la densité de $X_T$ et ses dérivées par celle $X_T^n$. Afin d'obtenir ce résultat, nous emploierons une méthode de calcul de Malliavin adaptatif basée sur les variables aléatoires utilisées dans le schéma. L'intérêt de notre approche repose sur le fait que l'on ne traite pas le cas d'un schéma particulier. Ainsi notre résultat s'applique aussi bien aux schémas d'Euler ($h=1$) que de Ninomiya Victoir ($h=2$) mais aussi à un ensemble générique de schémas. De plus les variables aléatoires utilisées dans le schéma n'ont pas de lois de probabilité imposées mais appartiennent à un ensemble de lois ce qui conduit à considérer notre résultat comme un principe d'invariance. On illustrera également ce résultat dans le cas d'un schéma d'ordre 3 pour les EDSs unidimensionnelles. La deuxième partie de cette thèse traite le sujet de l'estimation des paramètres d'une EDS. Ici, on va se placer dans le cas particulier de l'Estimateur du Maximum de Vraisemblance (EMV) des paramètres qui apparaissent dans le modèle matriciel de Wishart. Ce processus est la version multi-dimensionnelle du processus de Cox Ingersoll Ross (CIR) et a pour particularité la présence de la fonction racine carrée dans le coefficient de diffusion. Ainsi ce modèle permet de généraliser le modèle d'Heston au cas d'une covariance locale. Dans cette thèse nous construisons l'EMV des paramètres du Wishart. On donne également la vitesse de convergence et la loi limite pour le cas ergodique ainsi que pour certains cas non ergodiques. Afin de prouver ces convergences, nous emploierons diverses méthodes, en l'occurrence : les théorèmes ergodiques, des méthodes de changement de temps, ou l'étude de la transformée de Laplace jointe du Wishart et de sa moyenne. De plus, dans dernière cette étude, on étend le domaine de définition de cette transformée jointe / The development of technology and computer science in the last decades, has led the emergence of numerical methods for the approximation of Stochastic Differential Equations (SDE) and for the estimation of their parameters. This thesis treats both of these two aspects. In particular, we study the effectiveness of those methods. The first part will be devoted to SDE's approximation by numerical schemes while the second part will deal with the estimation of the parameters of the Wishart process. First, we focus on approximation schemes for SDE's. We will treat schemes which are defined on a time grid with size $n$. We say that the scheme $ X^n $ converges weakly to the diffusion $ X $, with order $ h in mathbb{N} $, if for every $ T> 0 $, $ vert mathbb{E} [f (X_T) -f (X_T^n)]vert leqslant C_f / h^n $. Until now, except in some particular cases (Euler and Victoir Ninomiya schemes), researches on this topic require that $ C_f$ depends on the supremum norm of $ f $ as well as its derivatives. In other words $C_f =C sum_{vert alpha vert leqslant q} Vert partial_{alpha} f Vert_{ infty}$. Our goal is to show that, if the scheme converges weakly with order $ h $ for such $C_f$, then, under non degeneracy and regularity assumptions, we can obtain the same result with $ C_f=C Vert f Vert_{infty}$. We are thus able to estimate $mathbb{E} [f (X_T)]$ for a bounded and measurable function $f$. We will say that the scheme converges for the total variation distance, with rate $h$. We will also prove that the density of $X^n_T$ and its derivatives converge toward the ones of $X_T$. The proof of those results relies on a variant of the Malliavin calculus based on the noise of the random variable involved in the scheme. The great benefit of our approach is that it does not treat the case of a particular scheme and it can be used for many schemes. For instance, our result applies to both Euler $(h = 1)$ and Ninomiya Victoir $(h = 2)$ schemes. Furthermore, the random variables used in this set of schemes do not have a particular distribution law but belong to a set of laws. This leads to consider our result as an invariance principle as well. Finally, we will also illustrate this result for a third weak order scheme for one dimensional SDE's. The second part of this thesis deals with the topic of SDE's parameter estimation. More particularly, we will study the Maximum Likelihood Estimator (MLE) of the parameters that appear in the matrix model of Wishart. This process is the multi-dimensional version of the Cox Ingersoll Ross (CIR) process. Its specificity relies on the square root term which appears in the diffusion coefficient. Using those processes, it is possible to generalize the Heston model for the case of a local covariance. This thesis provides the calculation of the EMV of the parameters of the Wishart process. It also gives the speed of convergence and the limit laws for the ergodic cases and for some non-ergodic case. In order to obtain those results, we will use various methods, namely: the ergodic theorems, time change methods or the study of the joint Laplace transform of the Wishart process together with its average process. Moreover, in this latter study, we extend the domain of definition of this joint Laplace transform

>Magister Scientiae - MSc / Herbal medicines have been an important part of health and wellness for hundreds of years. Recently the World Health Organization estimated that 80% of people worldwide rely on herbal medicines. Herbs contain many substances that are good for protecting the body and are therefore used in the treatment of various illnesses. Along with traditional medicines, herbs are often used in the treatment of chronic diseases such as rheumatism, migraine, chronic fatigue, asthma, eczema, and irritable bowel syndrome, among others. Herbal medicines are also applied in certain traditional communities as treatment against infectious diseases such as flu, malaria, measles, and even human immunodeficiency virus HIV-infection. Approximately 34 million people are currently infected with the human immunodeficiency virus (HIV) and 2.5 million newly infected. Therefore, HIV has become one of the major public health problems worldwide. It is important to understand the impact of herbal medicines used on HIV/AIDS. Mathematical models enable us to make predictions about the qualitative behaviour of disease outbreaks and evaluation of the impact of prevention or intervention strategies. In this dissertation we explore mathematical models for studying the effect of usage of herbal medicines on HIV. In particular we analyze a mathematical model for population dynamics of HIV/AIDS. The latter will include the impact of herbal medicines and traditional healing methods. The HIV model exhibits two steady states; a trivial steady state (HIV-infection free population) and a non-trivial steady state (persistence of HIV infection). We investigate the local asymptotic stability of the deterministic epidemic model and similar properties in terms of the basic reproduction number. Furthermore, we investigate for optimal control strategies. We study a stochastic version of the deterministic model by introducing white noise and show that this model has a unique global positive solution. We also study computationally the stochastic stability of the white noise perturbation model. Finally, qualitative results are illustrated by means of numerical simulations. Some articles from the literature that feature prominently in this dissertation are [14] of Cai et al, [10] of Bhunu et al., [86] of Van den Driessche and Watmough, [64] of Naresh et al., Through the study in this dissertation, we have prepared a research paper [1], jointly with the supervisors to be submitted for publication in an accredited journal. The author of this dissertation also contributed to the research paper [2], which close to completion. 1. Abdulaziz Y.A. Mukhtar, Peter J. Witbooi and Gail D. Hughes. A mathematical model for population dynamics of HIV with ARV and herbal medicine. 2. P.J. Witbooi, T. Seatlhodi, A.Y.A. Mukhtar, E. Mwambene. Mathematical modeling of HIV/AIDS with recruitment of infecteds.

Cette thèse concerne l’analyse et le développement de schémas d’intégration numérique de la Dynamique des Particules Dissipatives. Une présentation et une analyse de convergence faible de schémas existants est présentée, suivie d’une présentation et d’une analyse similaire de deux nouveaux schémas d’intégration facilement parallélisables. Une analyse des propriétés de conservation d’énergie de tous ces schémas est effectuée suivie d’une étude comparative de leurs biais sur l’estimation des valeurs moyennes d’observables physiques pour des systèmes à l’équilibre. Les schémas sont ensuite testés sur des systèmes choqués de fluides DPDE, où l’on montre que nos deux nouveaux schémas apportent une amélioration dans la précision de la description du comportement de tels systèmes par rapport aux schémas facilement parallélisables existants.Finalement, nous présentons une tentative d’accélération d’un schéma d’intégration de référence s’appliquant aux simulations séquentielles de la DPDE / This thesis is about the development and analysis of numerical schemes forthe integration of the Dissipative Particle Dynamics with Energy conservation. A presentation and a weak convergence analysis of existing schemes is performed, as well as the introduction and a similar analysis of two new straightforwardly parallelizable schemes. The energy preservation properties of all these schemes are studied followed by a comparative study of their biases on the estimation of the average values of physical observables on equilibrium simulations. The schemes are then tested on shock simulations of DPDE fluids, where we show that our schemes bring an improvement on the accuracy of the description of the behavior of such systems compared to existing straightforwardly parallelizable schemes. Finally, we present an attempt at accelerating a reference DPDE integration scheme on sequential simulations

In the present work, we study the topic of stochastic differential equations, their numerical solution and solution of models for pricing of options which follow from stochastic differential equations using the Itô calculus. We present several numerical methods for solving stochastic differential equations. These methods are then implemented in MATLAB and we investigate their properties, especially their convergence characteristics. Furthermore, we formulate two models for pricing of European call options. We solve these models using a variant of the spectral collocation method, again in MATLAB.

This thesis consists of four parts. In the first part we recall some background theory that will be used throughout the thesis. In the second part, we studied the absolute continuity of the laws of the solutions of some perturbed stochastic differential equaitons(SDEs) and perturbed reflected SDEs using Malliavin calculus. Because the extra terms in the perturbed SDEs involve the maximum of the solution itself, the Malliavin differentiability of the solutions becomes very delicate. In the third part, we studied the absolute continuity of the laws of the solutions of the parabolic stochastic partial differential equations(SPDEs) with two reflecting walls using Malliavin calculus. Our study is based on Yang and Zhang \cite{YZ1}, in which the existence and uniqueness of the solutions of such SPDEs was established. In the fourth part, we gave the existence and uniqueness of the solutions of the elliptic SPDEs with two reflecting walls and general diffusion coefficients.

Dans cette thèse, on s'intéresse à l'existence et à l'unicité d'une solution pour l'équation KPZ généralisée. On utilise la théorie récente des structures de régularité inspirée des chemins rugueux et introduite par Martin Hairer afin de donner sens à ce type d'équations singulières. La procédure de résolution comporte une partie algébrique à travers la définition du groupe de renormalisation et une partie stochastique avec la convergence de processus stochastiques renormalisés. Une des améliorations notoire de ce travail apportée aux structures de régularité est la définition du groupe de renormalisation par le biais d'une algèbre de Hopf sur des arbres labellés. Cette nouvelle construction permet d'obtenir des formules simples pour les processus stochastiques renormalisés. Ensuite, la convergence est obtenue par un traitement efficace de diagrammes de Feynman. / In this thesis, we investigate the existence and the uniqueness of the solution of the generalised KPZ equation. We use the recent theory of regularity structures inspired from the rough path and introduced by Martin Hairer in order to give a meaning to this singular equation. The procedure contains an algebraic part through the renormalisation group and a stochastic part with the computation of renormalised stochastic processes. One major improvement in the theory of the regularity structures is the definition of the renormalisation group using a Hopf algebra on some labelled trees. This new construction paves the way to simple formulas very useful for the renormalised stochastic processes. Then the convergence is obtained by an efficient treatment of some Feynman diagrams.

A standard problem in mathematical finance is the calculation of the price of some financial derivative such as various types of options. Since there exists analytical solutions in only a few cases it will often boil down to estimating the price with Monte Carlo simulation in conjunction with some numerical discretization scheme. The upside of using what we can call standard Monte Carlo is that it is relative straightforward to apply and can be used for a wide variety of problems. The downside is that it has a relatively slow convergence which means that the computational cost or complexity can be very large. However, this slow convergence can be improved upon by using Multilevel Monte Carlo instead of standard Monte Carlo. With this approach it is possible to reduce the computational complexity and cost of simulation considerably. The aim of this thesis is to introduce the reader to the Multilevel Monte Carlo method with applications to European and Asian call options in both the Black-Scholes-Merton (BSM) model and in the Heston model. To this end we first cover the necessary background material such as basic probability theory, estimators and some of their properties, the stochastic integral, stochastic processes and Ito’s theorem. We introduce stochastic differential equations and two numerical discretizations schemes, the Euler–Maruyama scheme and the Milstein scheme. We define strong and weak convergence and illustrate these concepts with examples. We also describe the standard Monte Carlo method and then the theory and implementation of Multilevel Monte Carlo. In the applications part we perform numerical experiments where we compare standard Monte Carlo to Multilevel Monte Carlo in conjunction with the Euler–Maruyama scheme and Milsteins scheme. In the case of a European call in the BSM model, using the Euler–Maruyama scheme, we achieved a cost O(ε-2(log ε)2) to reach the desired error in accordance with theory in comparison to the O(ε-3) cost for standard Monte Carlo. When using Milsteins scheme instead of the Euler–Maruyama scheme it was possible to reduce the cost in terms of the number of simulations needed to achieve the desired error even further. By using Milsteins scheme, a method with greater order of strong convergence than Euler–Maruyama, we achieved the O(ε-2) cost predicted by the complexity theorem compared to the standard Monte Carlo cost of order O(ε-3). In the final numerical experiment we applied the Multilevel Monte Carlo method together with the Euler–Maruyama scheme to an Asian call in the Heston model. In this case, where the coefficients of the Heston model do not satisfy a global Lipschitz condition, the study of strong or weak convergence is much harder. The numerical experiments suggested that the strong convergence was slightly slower compared to what was found in the case of a European call in the BSM model. Nevertheless, we still achieved substantial savings in computational cost compared to using standard Monte Carlo.

In the Thesis we study the problem of linear filtration of Gaussian signals in finite-dimensional space. We use the Kalman-type equations for the filter to show that the filter depends continuously on the signal. Secondly, we show the same continuity property for the covariance of the error and verify existence and uniqueness of a solution to an integral equation that is satisfied by the filter even under more general assumptions. We present several examples of application of the continuity property that are based on the theory of stochastic differential equations driven by fractional Brownian motion. 1