Dissertations / Theses on the topic 'Système d'EDP'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 27 dissertations / theses for your research on the topic 'Système d'EDP.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
Binard, Julie. "Modélisation, analyse et simulation de modèles en géomorphologie." Electronic Thesis or Diss., Université de Toulouse (2023-....), 2024. http://www.theses.fr/2024TLSEI001.
The study of landscape formation and evolution is a fundamental question in geomorphology. The work carried out in this thesis is devoted to the mathematical study of a landscape evolution model. Landform evolves in time due to erosion and sedimentation, caused by the flow of water over the ground surface. In particular, one of the objectives is to identify the processes involved in the formation of channels and patterns on the ground surface.In the introductory chapter, the model is derived from physical laws. It is a system composed of three partial differential equations, describing the fluid height, the concentration of sediments, and the surface elevation. In the second chapter, a mathematical study of the system is conducted. The existence and uniqueness of solutions to this system is proved. A study of the spectral stability of some stationary solutions is conducted, which allows determining the role of parameters in the appearance of patterns in the soil and providing information about their form. In particular, a necessary and sufficient condition for stability at all frequencies is given. The third chapter introduces a numerical particle method, for a class of stationary hyperbolic equations. The convergence of the numerical scheme is proven, and an error estimate is given. This stationary particle method is then applied to solve the equations for water height and sediment concentration, when water is in a stationary regime
Bréhier, Charles-Edouard. "Analyse numérique d'EDP Stochastiques hautement oscillantes." Phd thesis, École normale supérieure de Cachan - ENS Cachan, 2012. http://tel.archives-ouvertes.fr/tel-00763340.
Fougeirol, Jérémie. "Structure de variété de Hilbert et masse sur l'ensemble des données initiales relativistes faiblement asymptotiquement hyperboliques." Thesis, Avignon, 2017. http://www.theses.fr/2017AVIG0417/document.
General relativity is a gravitational theory born a century ago, in which the universe is a 4-dimensional Lorentzian manifold (N,gamma) called spacetime and satisfying Einstein's field equations. When we separate the time dimension from the three spatial ones, constraint equations naturally follow on from the 3+1 décomposition of Einstein's equations. Constraint equations constitute a necessary condition,as well as sufficient, to consider the spacetime N as the time evolution of a Riemannian hypersurface (m,g) embeded into N with the second fundamental form K. (m,g,K) is then an element of C, the set of initial data solutions to the constraint equations. In this work, we use Robert Bartnik's method to provide a Hilbert submanifold structure on C for weakly asymptotically hyperbolic initial data, whose regularity can be related to the bounded L^{2} curvature conjecture. Difficulties arising from the weakly AH case led us to introduce two second order differential operators and we obtain Poincaré and Korn-type estimates for them. Once the Hilbert structure is properly described, we define a mass functional smooth on the submanifold C and compatible with our weak regularity assumptions. The geometrical invariance of the mass is studied and proven, only up to a weak regularity conjecture about coordinate changes near infinity. Finally, we make a correspondance between critical points of the mass and static metrics
Filali, Siham. "Application du calcul stochastique à une classe d'EDP nonlinéaires." Lille 1, 2005. https://pepite-depot.univ-lille.fr/RESTREINT/Th_Num/2005/50376-2005-266.pdf.
Filali, Siham. "Application du calcul stochastique à une classe d'EDP non linéaire." Phd thesis, Université des Sciences et Technologie de Lille - Lille I, 2005. http://tel.archives-ouvertes.fr/tel-00012025.
obtenir l'existence et l'unicité de la solution d'un système d'équations aux
dérivées partielles non linéaire dont l'origine remonte à l'étude des modèles de particules collantes.
Premièrement, on construit deux diffusions dirigées par des browniens indépendants issues de points différents mais dont la dérive est la même fonction qui combine les deux densités de l'une et l'autre diffusions. On montre que le bonne combinaison de la densité et de la vitesse des particules est solution d'un système d'équations aux dérivées partielles appelé système de gaz sans pression avec viscosité.
Deuxièmement, On reprend la problématique d'un article de Sheu sur les densités de transition d'une diffusion non dégénéré, on aboutit à une meilleure précision sur les constantes apparaissant dans l'estimation de Sheu.
Finalement, on généralise le système de gaz sans pression déjà étudié par A. Dermoune en 2003, en remplaçant le laplacien par un opérateur plus générale. Alors on montre: l'existence d'une solution faible pour une équation différentielles stochastique non linéaire, identification de la dérive et l'unicité de la solution.
Takahashi, Takéo. "Analyse théorique, analyse numérique et contrôle de systèmes d'interaction fluide-structure et de systèmes de type ondes." Habilitation à diriger des recherches, Université Henri Poincaré - Nancy I, 2008. http://tel.archives-ouvertes.fr/tel-00590675.
Le, cavil Anthony. "Représentation probabiliste de type progressif d'EDP nonlinéaires nonconservatives et algorithmes particulaires." Thesis, Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLY023.
This thesis performs forward probabilistic representations of nonlinear and nonconservative Partial Differential Equations (PDEs), which allowto numerically estimate the corresponding solutions via an interacting particle system algorithm, mixing Monte-Carlo methods and non-parametric density estimates.In the literature, McKean typeNonlinear Stochastic Differential Equations (NLSDEs) constitute the microscopic modelof a class of PDEs which are conservative. The solution of a NLSDEis generally a couple $(Y,u)$ where $Y$ is a stochastic process solving a stochastic differential equation whose coefficients depend on $u$ and at each time $t$, $u(t,cdot)$ is the law density of the random variable $Y_t$.The main idea of this thesis is to consider this time a non-conservative PDE which is the result of a conservative PDE perturbed by a term of the type $Lambda(u, nabla u) u$. In this case, the solution of the corresponding NLSDE is again a couple $(Y,u)$, where again $Y$ is a stochastic processbut where the link between the function $u$ and $Y$ is more complicated and once fixed the law of $Y$, $u$ is determined by a fixed pointargument via an innovating Feynmann-Kac type formula
Zekraoui, Salim. "Contrôle et estimation en temps fini de certaines classes d'EDP." Electronic Thesis or Diss., Centrale Lille Institut, 2023. http://www.theses.fr/2023CLIL0028.
This Ph.D. thesis is devoted to the problems of non-asymptotic (finite, fixed, prescribed-time) estimation and stabilization of some classes of infinite-dimensional systems, namely LTI systems subject to input/sensor (pointwise or distributed) delays and reaction-diffusion PDEs. As the existing results on these classes of systems are few, we begin by reviewing relevant concepts and results on non-asymptotic tools (including homogeneity-based tools and time-varying tools) for finite-dimensional systems. Afterward, we extend these tools to infinite-dimensional settings. Firstly, we start with the problem of input and sensor delay compensation in finite/fixed/prescribed time of LTI systems where we use the so-called backstepping approach for PDEs (with some nonlinear and/or time-varying invertible transformations). To apply this approach, we reformulate the considered LTI system into a cascade ODE-PDE system where the PDE part is a hyperbolic transport equation that models the effect of the delay on the input/output. Secondly, we consider the problem of boundary state-dependent finite/fixed-time stabilization of reaction-diffusion PDEs. To the best of our knowledge, this problem has remained open in the literature for a considerable long time. We tackle this challenging problem using classical methods related to Control Lyapunov functions. We provide some hints on how we to extend this approach to input-to-state stabilization and non-asymptotic tracking problem for reaction-diffusion PDEs. We point out the limitations of our approach to observer design. Finally, we tackle the problem of input delay compensation of reaction-diffusion systems in prescribed time by output feedback using the backstepping approach. This problem is challenging, as one deals with observer and control designs with some time-varying gains that go to infinity when the time gets closer to the prescribed time of convergence, which brings additional challenges and issues. Dealing with these challenges requires introducing novel infinite-dimensional time-varying backstepping transformations in conjunction with advanced predictor-based concepts adapted to parabolic PDEs
Aoun, Mirella. "Analyse et analyse numérique d'EDP issues de la thermomécanique des fluides." Electronic Thesis or Diss., Normandie, 2023. http://www.theses.fr/2023NORMR093.
In this thesis, we focus on nonlinear evolutionary systems derived from a non-isothermal solidification problem with melt convection. These systems consist of three partial differential equations. The first is the phase-field equation coupled with the heat equation and the incompressible Navier-Stokes equation. More precisely, we are interested in the existence of solutions for these types of systems in the two-dimensional and the three-dimensional cases, and in the convergence of a finite volume approximation. One of the particularities of this type of system is the presence of a term naturally in L^1 in the energy conservation equation, which requires special treatment.This thesis is divided into two parts.The first part is divided into two chapters and is devoted to the study of problems with L^1 data and Neumann-type boundary conditions. To deal with these problems, and with data that are not very regular, we use the framework of renormalized solutions.In the first chapter, we establish a convergence result for solutions approximated by the finite volume method to the unique renormalized solution with zero median in the case of an elliptic convection-diffusion equation. In the second chapter, we focus on a non-linear parabolic problem with non-homogeneous Neumann conditions and a convection term. For this problem, we provide a definition of a renormalized solution and we show the existence and uniqueness of such a solution.The second part is devoted to the study of the system in dimensions 2 and 3. The first chapter deals with the dimension 2 and defines the notion of weak--renormalized solutions. With the help of the existence and stability results established in the first part for the conservation of energy equation, we prove the existence of a weak--renormalized solution.The final chapter considers the trickier case of dimension 3. The absence of a general stability and uniqueness result for the 3-dimensional Navier-Sokes equation requires us to transform the system into a formally equivalent one. By approximation and passage to the limit, we prove the existence of a solution in a weak sense
Le, cavil Anthony. "Représentation probabiliste de type progressif d'EDP nonlinéaires nonconservatives et algorithmes particulaires." Electronic Thesis or Diss., Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLY023.
This thesis performs forward probabilistic representations of nonlinear and nonconservative Partial Differential Equations (PDEs), which allowto numerically estimate the corresponding solutions via an interacting particle system algorithm, mixing Monte-Carlo methods and non-parametric density estimates.In the literature, McKean typeNonlinear Stochastic Differential Equations (NLSDEs) constitute the microscopic modelof a class of PDEs which are conservative. The solution of a NLSDEis generally a couple (Y,u) where Y is a stochastic process solving a stochastic differential equation whose coefficients depend on u and at each time t, u(t,.) is the law density of the random variable Yt.The main idea of this thesis is to consider this time a non-conservative PDE which is the result of a conservative PDE perturbed by a term of the type Lambda(u, nabla u) u. In this case, the solution of the corresponding NLSDE is again a couple (Y,u), where again Y is a stochastic processbut where the link between the function u and Y is more complicated and once fixed the law of Y, u is determined by a fixed pointargument via an innovating Feynmann-Kac type formula
Auriol, Jean. "Contrôle robuste d'EDPs linéaires hyperboliques par méthodes de backstepping." Thesis, Paris Sciences et Lettres (ComUE), 2018. http://www.theses.fr/2018PSLEM011/document.
Linear First-Order Hyperbolic Partial Differential Equations (LFOH PDEs) represent systems of conservation and balance law and are predominant in modeling of traffic flow, heat exchanger, open channel flow or multiphase flow. Different control approaches have been tackled for the stabilization or observation of such systems. Among them, the backstepping method consists to map the original system to a simpler system for which the control design is easier. The resulting controllers are explicit.In the first part of this thesis, we develop some general results in control theory. More precisely, we solve the problem of finite-time stabilization of a general class of LFOH PDEs using the backstepping methodology. The minimum stabilization time reachable may change depending on the number of available actuators. The corresponding boundary observers (crucial to envision industrial applications) are obtained through a dual approach. An important by-product of the proposed approach is to derive an explicit mapping from the space generated by the solutions of the considered LFOH PDEs to the space generated by the solutions of a general class of neutral systems with distributed delays. This mapping opens new prospects in terms of stability analysis for LFOH PDEs, extending the stability analysis methods developed for neutral systems.In the second part of the thesis, we prove the necessity of a change of strategy for robust control while considering industrial applications, for which the major limitation is known to be the robustness of the resulting control law to uncertainties in the parameters, delays in the loop, neglected dynamics or disturbances and noise acting on the system. In some situations, one may have to renounce to finite-time stabilization to ensure the existence of robustness margins. We propose some adjustments in the previously designed control laws by means of several degrees of freedom enabling trade-offs between performance and robustness. The robustness analysis is fulfilled using the explicit mapping between LFOH PDEs and neutral systems previously introduced
Lamothe, Vincent. "Transformations de Bäcklund, symétries et solutions explicites des systèmes d'EDPs." Thèse, Université du Québec à Trois-Rivières, 2005. http://depot-e.uqtr.ca/1984/1/000132219.pdf.
Bourgeau, Vanessa. "DÉVELOPEMENT D'UN SYSTÈME DE RÉGÉNÉRATION D'UDP-GA1NAC POUR LA GLYCOSYLATION ENZYMATIQUE D'OLIGOSACCHARIDES ET DE PEPTIDES D'INTÉRÊT THÉRAPEUTIQUE." Phd thesis, Université d'Orléans, 2006. http://tel.archives-ouvertes.fr/tel-00160999.
Nous avons mis au point un système de synthèse chimio enzymatique de glycoconjugués à Ga1NAc qui utilise 4 enzymes et des substrats simples comme le Ga1NAc, l'UTP et la créatine-P. Ce système permet la synthèse rapide et efficace de glycopeptides et d'oligosaccharides à GaINAc ; il a été utilisé pour glycosyler l'antigène MUC1 et nous avons pu évaluer la réponse immunitaire murine contre la petite glycoprotéine obtenue.
Le système de synthèse d'UDP-Ga1NAc peut accepter certains dérivés de Ga1NAc et permettre ainsi la synthèse d'analogues d'UDP-Ga1NAc. Ces molécules sont des sondes appréciables pour étudier l'interaction entre substrats et enzymes par les techniques physicochimiques comme la STDNMR.
Bourgeaux, Vanessa. "Développement d'un système de régénération d'UDP-GalNAc pour la glycosylation enzymatique d'oligosaccharides et de peptides d'intérêt thérapeutique." Orléans, 2006. http://www.theses.fr/2006ORLE2049.
Chenikhar, Karim. "Evaluation d'outils de preuves formelles pour la specification et la validation des comportements des systemes de controle-commande d'edf." Paris, CNAM, 2000. http://www.theses.fr/2001CNAM0367.
Khemmoudj, Mohand Ou Idir. "Modélisation et résolution de systèmes de contraintes : application au problème de placement des arrêts et de la production des réacteurs nucléaires d'EDF." Paris 13, 2007. http://www.theses.fr/2007PA132010.
Diedhiou, Moussa Mory. "Approche mixte interface nette-diffuse pour les problèmes d'intrusion saline en sous-sol : modélisation, analyse mathématique et illustrations numériques." Thesis, La Rochelle, 2015. http://www.theses.fr/2015LAROS023/document.
The context of the subject is the management of aquifers, in especially the control of their operations and their possible pollution. A critical case is the saltwater intrusion problem in costal aquifers. The goal is to obtain efficient and accurate models to simulate the displacement of fresh and salt water fronts in coastal aquifer for the optimal exploitation of groundwater. More generally, the work applies for miscible and stratified displacements in slightly deformable porous media. In this work we propose an original model mixing abrupt interfaces/diffuse interfaces approaches. The advantage is to adopt the (numerical) simplicity of a sharp interface approach, and to take into account the existence of diffuse interfaces. The model is based on the conservation laws written in the saltwater zone and in the freshwater zone, these two free boundary problems being coupled through an intermediate phase field model. An upscaling procedure let us reduce the problem to a two-dimensional setting. The theoretical analysis of the new model is performed. We also present numerical simulations comparing our 2D model with the classical 3D model for miscible displacement in a confined aquifer. Physical predictions from our new model are also given for an unconfined setting
Lemaire, Vincent. "Estimation récursive de la mesure invariante d'un processus de diffusion." Phd thesis, Université de Marne la Vallée, 2005. http://tel.archives-ouvertes.fr/tel-00011281.
La principale hypothèse sur ces solutions (diffusions) est l'existence d'une fonction de Lyapounov garantissant une condition de stabilité. Par le théorème ergodique on sait que les mesures empiriques de la diffusion convergent vers une mesure invariante. Nous étudions une convergence similaire lorsque la diffusion est discrétisée par un schéma d'Euler de pas décroissant. Nous prouvons que les mesures empiriques pondérées de ce schéma convergent vers la mesure invariante de la diffusion, et qu'il est possible d'intégrer des fonctions exponentielles lorsque le coefficient de diffusion est suffisamment petit. De plus, pour une classe de diffusions plus restreinte, nous prouvons la convergence presque sûre et dans Lp du schéma d'Euler vers la diffusion.
Nous obtenons des vitesses de convergence pour les mesures empiriques pondérées et donnons les paramètres permettant une vitesse optimale. Nous finissons l'étude de ce schéma lorsqu'il y a présence de multiples mesures invariantes. Cette étude se fait en dimension 1, et nous permet de mettre en évidence un lien entre classification de Feller et fonctions de Lyapounov.
Dans la dernière partie, nous exposons un nouvel algorithme adaptatif permettant de considérer des problèmes plus généraux tels que les systèmes Hamiltoniens ou les systèmes monotones. Il s'agit de considérer les mesures empiriques d'un schéma d'Euler construit à partir d'une suite de pas aléatoires adaptés dominée par une suite décroissant vers 0.
Boussaada, Islam. "Contribution à l'étude des solutions périodiques et des centres isochrones des systèmes d'équations différentielles ordinaires plans." Phd thesis, Rouen, 2008. http://www.theses.fr/2008ROUES056.
The first part (which is an already published paper, written in collaboration with R. Chouikha) is devoted to the search of periodic solutions of "generalized Liénard equation". A theorem is proved which insures the existence of such solutions under appropriate assumptions. The second part is devoted to the search of isochronous centers of the planar polynomial systems of ordinary differential equations. Using C-algorithm we determine eight new cases of isochronous centers. We prove also the efficiency of the normal forms method for such investigations ; studying some systems of order 2, 3, 4 and recovering in uniform way some already known results
Boussaada, Islam. "Contribution à l'étude des solutions périodiques et des centres isochrones des systèmes d'équations différentielles ordinaires plans." Phd thesis, Université de Rouen, 2008. http://tel.archives-ouvertes.fr/tel-00348281.
La première partie, (il s'agit d'un travail publié et écrit en collaboration avec R. Chouikha) est consacré à la recherche des solutions périodiques de « l'équation de Liénard généralisée ». On démontre un théorème qui asure dans certains cas l'existence de telles solutions.
La seconde partie est consacré à la recherche de centres isochrones de systèmes d'équations différentielles ordinaires polynomiaux plans. Grâce à l'usage de C-algorithme, on détermine huit nouveaux cas. On montre aussi l'efficacité de la méthode des formes normales dans de telles recherches, en examinant des systèmes d'ordre 2, 3, 4 et en retrouvant de manière uniforme plusieurs résultats déjà connus.
Barril, Basil Carles. "Semilinear hyperbolic equations and the dynamics of gut bacteria." Doctoral thesis, Universitat Autònoma de Barcelona, 2018. http://hdl.handle.net/10803/643304.
In this thesis we propose a mathematical framework to analyse the dynamics of microorganisms growing within the guts of animals. Such a framework consists of a hyperbolic system of PDEs with non-linear reaction terms and certain boundary conditions that link the microbes in the environment with those inside the hosts. In chapter 1 we solve the Abstract Cauchy Problem associated to the model by considering the semilinear formulation on a certain Banach space X. The semilinear structure of the system obtained in this way is special because, on the one hand, the evolution law can be expressed as the sum of a linear unbounded operator and a non-linear Lipschitz function (which is typical) but, on the other hand, the non-linear perturbation takes values not in X but on a larger space Y which is related to X (which is atypical). In order to deal with this situation we use the theory of dual semigroups. Stability results around steady states are also given when the nonlinear perturbation is Fréchet differentiable. These results are based on two propositions: one relating the local dynamics of the non-linear semiow with the linearised semigroup around the equilibrium, and a second relating the dynamical properties of the linearised semigroup with the spectral values of its generator. The later is proven by showing that the Spectral Mapping Theorem always applies to the semigroups one obtains when the semiow is linearised. In chapter 2 an autonomous semi-linear hyperbolic pde system for the proliferation of bacteria within a heterogeneous population of animals is presented and analysed. It is assumed that bacteria grow inside the intestines and that they can be either attached to the epithelial wall or as free particles in the lumen. A condition involving ecological parameters is given, which can be used to decide the existence of endemic equilibria as well as local stability properties of the non-endemic one. Some implications on phage therapy are addressed. In chapter 3 the basic reproduction number associated to the bacterial population, i.e. the expected number of daughter cells per bacterium, is given explicitly in terms of biological parameters. In addition, an alternative quantity is introduced based on the number of bacteria produced within the intestine by one bacterium originally in the external media. The latter depends on the parameters in a simpler way and provides more biological insight than the standard reproduction number, allowing the design of experimental procedures. Both quantities coincide and are equal to one at the extinction threshold, below which the bacterial population becomes extinct. Optimal values of both reproduction numbers are derived assuming parameter trade-offs.
Li, Ji. "Analyse mathématique de modèles d'intrusion marine dans les aquifères côtiers." Thesis, Littoral, 2015. http://www.theses.fr/2015DUNK0378/document.
The theme of this thesis is the analysis of mathematical models describing saltwater intrusion in coastal aquifers. The simplicity of sharp interface approach is chosen : there is no mass transfer between fresh water and salt water (respectively between the saturated zone and the area dry). We compensate the mathematical difficulty of the analysis of free interfaces by a vertical averaging process allowing us to reduce the 3D problem to system of pde's defined on a 2D domain Ω. A second model is obtained by combining the approach of 'sharp interface' in that with 'diffuse interface' ; this approach is derived from the theory introduced by Allen-Cahn, using phase functions to describe the phenomena of transition between fresh water and salt water (respectively the saturated and unsaturated areas). The 3D problem is then reduced to a strongly coupled system of quasi-linear parabolic equations in the unconfined case describing the evolution of the DEPTHS of two free surfaces and elliptical-parabolic equations in the case of confined aquifer, the unknowns being the depth of salt water/fresh water interface and the fresh water hydraulic head. In the first part of the thesis, the results of global in time existence are demonstrated showing that the sharp-diffuse interface approach is more relevant since it allows to establish a mor physical maximum principle (more precisely a hierarchy between the two free surfaces). In contrast, in the case of confined aquifer, we show that both approach leads to similar results. In the second part of the thesis, we prove the uniqueness of the solution in the non-degenerate case. The proof is based on a regularity result of the gradient of the solution in the space Lr (ΩT), r > 2, (ΩT = (0,T) x Ω). Then we are interest in a problem of identification of hydraulic conductivities in the unsteady case. This problem is formulated by an optimization problem whose cost function measures the squared difference between experimental hydraulic heads and those given by the model
Penel, Yohan. "Etude théorique et numérique de la déformation d'une interface séparant deux fluides non-miscibles à bas nombre de Mach." Phd thesis, Université Paris-Nord - Paris XIII, 2010. http://tel.archives-ouvertes.fr/tel-00547865.
Forestier-Coste, Louis. "Croissance et coalescence de bulles dans les magmas : analyse mathématique et simulation numérique." Phd thesis, Université d'Orléans, 2012. http://tel.archives-ouvertes.fr/tel-00736634.
Ley, Olivier. "Evolution de fronts avec vitesse non-locale et équations de Hamilton-Jacobi." Habilitation à diriger des recherches, Université François Rabelais - Tours, 2008. http://tel.archives-ouvertes.fr/tel-00362409.
Le premier chapitre concerne l'évolution de fronts avec une vitesse normale prescrite. Pour étudier ce genre de problème, une première approche, dite par lignes de niveaux, consiste àreprésenter le front comme une ligne de niveau d'une fonction auxiliaire u. Cette approche ramène l'étude du problème d'évolution géométrique à un problème d'EDP puisque u vérifie une équation de Hamilton-Jacobi. Quelques résultats dans le cas de vitesses locales comme la courbure moyenne sont présentés mais la majorité des résultats concerne le cas de vitesses non-locales décrivant la dynamique des dislocations dans un cristal ou modélisant l'asymptotique d'un système de FitzHugh-Nagumo apparaissant en biologie. Une approche différente, basée sur des solutions de viscosité géométriques, est utilisée pour étudier des problèmes de propagation de fronts apparaissant en optimisation de formes. Le but est de trouver un ensemble optimal minimisant une énergie du type capacité à volume ou périmètre constant. L'idée est de déformer le bord d'un ensemble donné avec une vitesse normale adéquate de manière à diminuer au plus son énergie. La mise en oeuvre de cette idée nécessite la construction rigoureuse d'une telle évolution pour tout temps et la preuve de la convergence vers une solution du problème initial. De plus, la décroissance de l'énergie est obtenue le long du flot.
Le deuxième chapitre décrit des résultats d'unicité, d'existence et d'homogénéisation pour des équations de Hamilton-Jacobi-Bellman. La majeure partie du travail effectué concerne des équations provenant de problèmes de contrôle stochastique avec des contrôles non-bornés. Les équations comportent alors des termes quadratiques par rapport au gradient et les solutions étudiées sont elles-mêmes à croissance quadratique. Des liens entre ces solutions et les fonctions valeurs des problèmes de contrôle correspondants sont établis. La seconde partie est consacrée à un théorème d'homogénéisation pour un système d'équations de Hamilton-Jacobi du premier ordre.
Le troisième et dernier chapitre traite d'un sujet un peu à part, à savoir le lien entre les flots de gradient et l'inégalité de Lojasiewicz. La principale originalité de ce travail est de placer l'étude dans un cadre hilbertien pour des fonctions semiconvexes, ce qui sort du cadre de l'inégalité de Lojasiewicz classique. Le principal théorème produit des caractérisations de cette inégalité. Les résultats peuvent être précisés dans le cas des fonctions convexes ; en particulier, un contre-exemple de fonction convexe ne vérifiant pas l'inégalité de Lojasiewicz est construit. Cette dernière inégalité est reliée à la longueur des trajectoires de gradient. Une borne de cette longueur est obtenue pour les fonctions convexes coercives en dimension deux même lorsque cette inégalité n'est pas vérifiée.
Benosman, Chahrazed. "Contrôle de la Dynamique de la Leucémie Myéloïde Chronique par Imatinib." Phd thesis, Bordeaux 1, 2010. http://tel.archives-ouvertes.fr/tel-00555973.
Rousset, Mathias. "Méthodes de "Population Monte-Carlo'' en temps continu est physique numérique." Toulouse 3, 2006. http://www.theses.fr/2006TOU30251.
In this dissertation, we focus on stochastic numerical methods of Population Monte-Carlo type, in the continuous time setting. These PMC methods resort to the sequential computation of averages of weighted Markovian paths. The practical implementation rely then on the time evolution of the empirical distribution of a system of N interacting walkers. We prove the long time convergence (towards Schrödinger groundstates) of the variance and bias of this method with the expected 1/N rate. Next, we consider the problem of sequential sampling of a continuous flow of Boltzmann measures. For this purpose, starting with any Markovian dynamics, we associate a second dynamics in reversed time whose law (weighted by a computable Feynman-Kac path average) gives out the original dynamics as well as the target Boltzmann measure. Finally, we generalize the latter problem to the case where the dynamics is caused by evolving rigid constraints on the positions of the process. We compute exactly the associated weights, which resorts to the local curvature of the manifold defined by the constraints