This work is devoted to the theoretical and numerical aspects of shape optimization. The first part (chapter I to IV) deals with optimization problems under convexity constraint or constant width constraint. We give several new results related to Newton's problem and Meissner's conjecture. The second part (chapter V to VII) deals with the numerical study of shape optimization problems where many shapes or phases are involved. Some new numerical methods are introduced to study optimal configurations of famous problems : Kelvin's problem and Caffarelli's conjecture. The last part (chapter VIII and IX) is devoted to optimal transportation problems and irrigation problems. More precisely, we introduce a general framework, where different kind of cost functions are allowed. This seems relevant in some problems presenting congestion effects as for instance traffic on a highway, crowds moving in domains with obstacles. In the last chapter we give preliminary results related to the numerical approximation of optimal irrigation networks.

This thesis is devoted to subriemannian optimal transportation problems. In the first part of the thesis, we consider cost functions arising from very general optimal control costs. We prove the existence and uniqueness of an optimal map between two given measures under certain regularity and growth assumptions on the Lagrangian, absolute continuity of the measures with respect to the Lebesgue class, and, most importantly, the absence of sharp abnormal minimizers. In particular, this result is applicable in the case where the cost function is square of the subriemannian distance on a subriemannian manifold with a 2-generating distribution. This unifies and generalizes the corresponding Riemannian and subriemannian results of Brenier, McCann, Ambrosio-Rigot and Bernard-Buffoni. We also establish various properties of the optimal plan when abnormal minimizers are present. The second part of the thesis is devoted to the infinite-dimensional geometry of optimal transportation on a subriemannian manifold. We start by proving the following nonholonomic version of the classical Moser theorem: given a bracket-generating distribution on a connected compact manifold (possibly with boundary), two volume forms of equal total volume can be isotoped by the flow of a vector field tangent to this distribution. Next, we describe formal solutions of the corresponding subriemannian optimal transportation problem and present the Hamiltonian framework for both the Otto calculus and its subriemannian counterpart as infinite-dimensional Hamiltonian reductions on diffeomorphism groups. Finally, we define a subriemannian analog of the Wasserstein metric on the space of densities and prove that the subriemannian heat equation defines a gradient flow on the subriemannian Wasserstein space with the potential given by the Boltzmann relative entropy functional. Measure contraction property is one of the possible generalizations of Ricci curvature bound to more general metric measure spaces. In the third part of the thesis, we discuss when a three dimensional contact subriemannian manifold satisfies such property.

This thesis is devoted to subriemannian optimal transportation problems. In the first part of the thesis, we consider cost functions arising from very general optimal control costs. We prove the existence and uniqueness of an optimal map between two given measures under certain regularity and growth assumptions on the Lagrangian, absolute continuity of the measures with respect to the Lebesgue class, and, most importantly, the absence of sharp abnormal minimizers. In particular, this result is applicable in the case where the cost function is square of the subriemannian distance on a subriemannian manifold with a 2-generating distribution. This unifies and generalizes the corresponding Riemannian and subriemannian results of Brenier, McCann, Ambrosio-Rigot and Bernard-Buffoni. We also establish various properties of the optimal plan when abnormal minimizers are present. The second part of the thesis is devoted to the infinite-dimensional geometry of optimal transportation on a subriemannian manifold. We start by proving the following nonholonomic version of the classical Moser theorem: given a bracket-generating distribution on a connected compact manifold (possibly with boundary), two volume forms of equal total volume can be isotoped by the flow of a vector field tangent to this distribution. Next, we describe formal solutions of the corresponding subriemannian optimal transportation problem and present the Hamiltonian framework for both the Otto calculus and its subriemannian counterpart as infinite-dimensional Hamiltonian reductions on diffeomorphism groups. Finally, we define a subriemannian analog of the Wasserstein metric on the space of densities and prove that the subriemannian heat equation defines a gradient flow on the subriemannian Wasserstein space with the potential given by the Boltzmann relative entropy functional. Measure contraction property is one of the possible generalizations of Ricci curvature bound to more general metric measure spaces. In the third part of the thesis, we discuss when a three dimensional contact subriemannian manifold satisfies such property.

In this thesis we study three different problems: convex ancient solutions to the power-of-mean curvature flow; Sharp inequalities; regularity results in some applications of optimal transportation. The second chapter is devoted to the power-of-mean curvature flow; We prove some estimates for convex ancient solutions (the existence time for the solution starts from -\infty) to the power-of-mean curvature flow, when the power is strictly greater than \frac{1}{2}. As an application, we prove that in two dimension, the blow-down of an entire convex translating solution, namely u_{h}=\frac{1}{h}u(h^{\frac{1}{1+\alpha}}x), locally uniformly converges to \frac{1}{1+\alpha}|x|^{1+\alpha} as h\rightarrow\infty. The second application is that for generalized curve shortening flow (convex curve evolving in its normal direction with speed equal to a power of its curvature), if the convex compact ancient solution sweeps the whole space \mathbb{R}^{2}, it must be a shrinking circle. Otherwise the solution must be defined in a strip region. In the first section of the third chapter, we prove a one-parameter family of sharp conformally invariant integral inequalities for functions on the $n$-dimensional unit ball. As a limiting case, we obtain an inequality that generalizes Carleman's inequality for harmonic functions in the plane to poly-harmonic functions in higher dimensions. The second section represents joint work with Tobias Weth and Rupert Frank; the main result is that, one can always put a sharp remainder term on the righthand side of the sharp fractional sobolev inequality. In the first section of the final chapter, under some suitable condition, we prove that the solution to the principal-agent problem must be C^{1}. The proof is based on a perturbation argument. The second section represents joint work with Emanuel Indrei; the main result is that, under (A3S) condition on the cost and c-convexity condition on the domains, the free boundary in the optimal partial transport problem is C^{1,\alpha}.

In this thesis we study three different problems: convex ancient solutions to the power-of-mean curvature flow; Sharp inequalities; regularity results in some applications of optimal transportation. The second chapter is devoted to the power-of-mean curvature flow; We prove some estimates for convex ancient solutions (the existence time for the solution starts from -\infty) to the power-of-mean curvature flow, when the power is strictly greater than \frac{1}{2}. As an application, we prove that in two dimension, the blow-down of an entire convex translating solution, namely u_{h}=\frac{1}{h}u(h^{\frac{1}{1+\alpha}}x), locally uniformly converges to \frac{1}{1+\alpha}|x|^{1+\alpha} as h\rightarrow\infty. The second application is that for generalized curve shortening flow (convex curve evolving in its normal direction with speed equal to a power of its curvature), if the convex compact ancient solution sweeps the whole space \mathbb{R}^{2}, it must be a shrinking circle. Otherwise the solution must be defined in a strip region. In the first section of the third chapter, we prove a one-parameter family of sharp conformally invariant integral inequalities for functions on the $n$-dimensional unit ball. As a limiting case, we obtain an inequality that generalizes Carleman's inequality for harmonic functions in the plane to poly-harmonic functions in higher dimensions. The second section represents joint work with Tobias Weth and Rupert Frank; the main result is that, one can always put a sharp remainder term on the righthand side of the sharp fractional sobolev inequality. In the first section of the final chapter, under some suitable condition, we prove that the solution to the principal-agent problem must be C^{1}. The proof is based on a perturbation argument. The second section represents joint work with Emanuel Indrei; the main result is that, under (A3S) condition on the cost and c-convexity condition on the domains, the free boundary in the optimal partial transport problem is C^{1,\alpha}.

Abordaremos a teoria do transporte otimo demonstrando o teorema da dualidade de Kantorovich para uma classe ampla de funções custo. Tal resultado desempenha um papel de suma importância na teoria do transporte otimo. Uma ferramenta importante utilizada e o teorema da dualidade de Fenchel-Rockafellar, aqui enunciado e demonstrado em bastante generalidade. Demonstramos tamb em o teorema da dualidade de Kantorovich-Rubinstein, que trata do caso particular da função custo distância. / We analyze the optimal transport theory proving the Kantorovich duality theorem for a wide class of cost functions. Such result plays an extremely important role in the optimal transport theory. An important tool used here is the Fenchel-Rockafellar duality theorem, which we state and prove in a general case. We also prove the Kantorovich-Rubinstein duality theorem, which deals with the particular case of cost function given by the distance.

Abordaremos a teoria do transporte otimo demonstrando o teorema da dualidade de Kantorovich para uma classe ampla de funções custo. Tal resultado desempenha um papel de suma importância na teoria do transporte otimo. Uma ferramenta importante utilizada e o teorema da dualidade de Fenchel-Rockafellar, aqui enunciado e demonstrado em bastante generalidade. Demonstramos tamb em o teorema da dualidade de Kantorovich-Rubinstein, que trata do caso particular da função custo distância. / We analyze the optimal transport theory proving the Kantorovich duality theorem for a wide class of cost functions. Such result plays an extremely important role in the optimal transport theory. An important tool used here is the Fenchel-Rockafellar duality theorem, which we state and prove in a general case. We also prove the Kantorovich-Rubinstein duality theorem, which deals with the particular case of cost function given by the distance.

Abordaremos a teoria do transporte otimo demonstrando o teorema da dualidade de Kantorovich para uma classe ampla de funções custo. Tal resultado desempenha um papel de suma importância na teoria do transporte otimo. Uma ferramenta importante utilizada e o teorema da dualidade de Fenchel-Rockafellar, aqui enunciado e demonstrado em bastante generalidade. Demonstramos tamb em o teorema da dualidade de Kantorovich-Rubinstein, que trata do caso particular da função custo distância. / We analyze the optimal transport theory proving the Kantorovich duality theorem for a wide class of cost functions. Such result plays an extremely important role in the optimal transport theory. An important tool used here is the Fenchel-Rockafellar duality theorem, which we state and prove in a general case. We also prove the Kantorovich-Rubinstein duality theorem, which deals with the particular case of cost function given by the distance.

Une part importante de l’activité en calcul scientifique et analyse numérique est consacrée aux problèmes de transport d’une quantité par un champ donné (ou lui-même calculé numériquement). Les questions de conservations étant essentielles dans ce domaine, on formule en général le problème de façon eulérienne sous la forme d’un bilan au niveau de chaque cellule élémentaire du maillage, et l’on gère l’évolution en suivant les valeurs moyennes dans ces cellules au cours du temps. Une autre approche consiste à suivre les caractéristiques du champ et à transporter les valeurs ponctuelles le long de ces caractéristiques. Cette approche est délicate à mettre en oeuvre, n’assure pas en général une parfaite conservation de la matière transportée, mais peut permettre dans certaines situations de transporter des quantités non régulières avec une grande précision, et sur des temps très longs (sans conditions restrictives sur le pas de temps comme dans le cas des méthodes eulériennes). Les travaux de thèse présentés ici partent de l’idée suivante : dans le cadre des méthodes utilisant un suivi de caractéristiques, transporter une quantité supplémentaire géométrique apportant plus d’informations sur le problème (on peut penser à un tenseur des contraintes dans le contexte de la mécanique des fluides, une métrique sous-jacente lors de l’adaptation de maillage, etc. ). Un premier pan du travail est la formulation théorique d’une méthode de transport de telles quantités. Elle repose sur le principe suivant : utiliser la différentielle du champ de transport pour calculer la différentielle du flot, nous donnant une information sur la déformation locale du domaine nous permettant de modifier nos quantités géométriques. Cette une approche a été explorée dans dans le contexte des méthodes particulaires plus particulièrement dans le domaine de la physique des plasmas. Ces premiers travaux amènent à travailler sur des densités paramétrées par un couple point/tenseur, comme les gaussiennes par exemple, qui sont un contexte d’applications assez naturelles de la méthode. En effet, on peut par la formulation établie transporter le point et le tenseur. La question qui se pose alors et qui constitue le second axe de notre travail est celle du choix d’une distance sur des espaces de densités, permettant par exemple d’étudier l’erreur commise entre la densité transportée et son approximation en fonction de la « concentration » au voisinage du point. On verra que les distances Lp montrent des limites par rapport au phénomène que nous souhaitons étudier. Cette étude repose principalement sur deux outils, les distances de Wasserstein, tirées de la théorie du transport optimal, et la distance de Fisher, au carrefour des statistiques et de la géométrie différentielle. / In applied mathematics, question of moving quantities by vector is an important question : fluid mechanics, kinetic theory… Using particle methods, we're going to move an additional quantity giving more information on the problem. First part of the work is the theorical formulation for this kind of transport. It's going to use the differential in space of the vector field to compute the differential of the flow. An immediate and natural application is density who are parametrized by and point and a tensor, like gaussians. We're going to move such densities by moving point and tensor. Natural question is now the accuracy of such approximation. It's second part of our work , which discuss of distance to estimate such type of densities.

Apresentamos o problema do transporte ótimo de Monge-Kantorovich com duas medidas de probabilidade conhecidas e que possuem suporte em um conjunto de cardinalidade finita. O objetivo é determinar condições que permitam construir um acoplamento destas medidas que minimiza o valor esperado de uma função de custo conhecida e que assume valor nulo apenas nos elementos da diagonal. Apresentamos também um resultado relacionado com a solução do problema de Monge-Kantorovich em espaços produto finitos quando conhecemos soluções para o problema nos espaços marginais. / We present the Monge-Kantorovich optimal problem with two known probability measures on a finite set. The objective is to obtain conditions that allow us to build a coupling of these measures that minimizes the expected value of a cost function that is known and is zero only on the diagonal elements. We also present a result that is related with the solution of the Monge-Kantorovich problem in finite product spaces in the case that solutions to the problem in the marginal spaces are known.