Neste trabalho analisamos a positividade definida estrita de núcleos contínuos sobre um espaço compacto 2-homogêneo. R. Gangolli (1967) apresentou uma caracterização completa para os núcleos que são contínuos, isotrópicos e positivos definidos sobre um espaço compacto 2-homogêneo Md: a parte isotrópica do núcleo é uma série de Fourier uniformemente convergente, com coeficientes não negativos, em relação a certos polinômios de Jacobi atrelados a Md. Uma das contribuições de nosso trabalho é uma caracterização para a positividade definida estrita de tais núcleos, complementando a caracterização apresentada por Chen et al. (2003) no caso em que Md é uma esfera unitária de dimensão maior ou igual a 2. Outra contribuição do trabalho é uma extensão do resultado de Gangolli para núcleos sobre produtos cartesianos de espaços compactos 2-homogêneos, e a consequente caracterização para núcleos estritamente positivos definidos neste mesmo contexto. Por fim, a última contribuição do trabalho envolve a análise do grau de diferenciabilidade da parte isotrópica de um núcleo contínuo, isotrópico e positivo definido sobre Md e a aplicabilidade de tal análise em resultados envolvendo a positividade definida estrita. / In this work we analyze the strict positive definiteness of continuous kernels on compact two-point homogeneous spaces Md. R. Gangolli (1967) presented a complete characterization for continuous, isotropic and positive definite kernels on Md: the isotropic part of the kernel is a uniformly convergent Fourier series of certain Jacobi polynomials associated to Md, with nonnegative coefficients. One of the contributions of our work is a characterization for the strict positive definiteness of such kernels, completing that one presented by Chen et al. (2003) in the case Md is the unit sphere of dimension at least 2. Another contribuition of this work is an extension of Gangolli\'s result for kernels on a product of compact two-point homogeneous spaces, and the subsequent characterization of strict positive definiteness in this same context. Finally, the last contribution in this work involves the analysis of the differentiability of the isotropic part of a continuous, isotropic and positive definite kernel on Md and the applicability of such analysis in results involving the strict positive definiteness.

Apresentamos uma caracterização das funções positivas definidas em esferas complexas, generalizando assim, um resultado de Schoenberg ([41]). Como no caso real, uma classe importante dessas funções é aquela composta pelas funções estritamente positivas definidas de uma certa ordem; estas podem ser utilizadas para resolver certos problemas de interpolação de dados arbitrários associados a pontos distintos distribuídos nas esferas. Com esse objetivo, obtivemos algumas condições necessárias e suficientes (separadamente) para que funções positivas definidas sejam estritamente positivas definidas. Os resultados apresentados fornecem uma caracterização final elementar para funções estritamente positivas definidas de todas as ordens em quase todas as esferas complexas. Funções estritamente positivas definidas de ordem 2 são caracterizadas em todas as esferas complexas. Analisamos também a relação entre funções estritamente positivas definidas em esferas complexas e funções estritamente positivas definidas em esferas reais. / We characterize positive definite functions on complex spheres, generalizing a famous result due to I. J. Schoenberg ([41]). As in the real case, we study the so-called strictly positive definite functions. They can be used to perform interpolation of scattered data on those spheres. We present (separated) necessary and sufficient conditions for a positive definite function to be strictly positive definite of a certain order. These conditions produce a final characterization for those positive definite functions which are strictly positive definite of all orders, on almost all spheres. Strictly positive definite functions of order 2 are identified. Finally, we study a connection between strictly positive definite functions on real spheres and strictly positive definite functions on complex spheres.

Esta dissertação aborda a teoria dos jogos diferenciais em sua estreita relação com a teoria das equações de Hamilton-Jacobi (HJ). Inicialmente, uma revisão da noção de solução em teoria dos jogos é empreendida. Discutem-se nesta ocasião as idéias de equilíbrio de Nash e alguns de seus refinamentos. Em seguida, tem lugar uma introdução à teoria dos jogos diferenciais, onde noções de solução como a função de valor de Isaacs e de Friedman são discutidas. É nesta altura do trabalho que fica evidente a conexão entre este conceito de solução e a teoria das equações de Hamilton-Jacobi. Por ocasião desta conexão, é explorada a noção de solução clássica e é exposta uma demonstração do fato de que se um jogo diferencial possuir uma função de valor pelo menos continuamente diferenciável, esta será uma solução da equação de Hamilton-Jacobi associada ao jogo. Este resultado faz uso do princípio da programação dinâmica, devido a Bellman, e cuja demonstração está presente no texto. No entanto, quando a função de valor do jogo é apenas contínua, então embora esta não seja uma solução clássica da equação HJ associada a jogo, vemos que ela será uma solução viscosa, ou solução no sentido da viscosidade - e a esta altura são discutidos os elementos e propriedades desta classe de soluções, um teorema de existência e unicidade e alguns exemplos. Por fim, retomamos o estudo dos jogos diferenciais à luz das soluções viscosas da equação de Hamilton-Jacobi e, assim, expomos uma demonstração de existência da função de valor e do princípio da programação dinâmica a partir das noções da viscosidade / This dissertation aims to address the topic of Differential Game Theory in its connection with the Hamilton-Jacobi (HJ) equations framework. Firstly we introduce the idea of solution for a game, through the discussion of Nash equilibria and its refinements. Secondly, the solution concept is then translated to the context of Differential Games and the idea of value function is introduced in its Isaacs\'s as well as Friedman\'s version. As the value function is discussed, its relationship with the Hamilton-Jacobi equations theory becomes self-evident. Due to such relation, we investigate the HJ equation from two distinct points of view. First of all, we discuss a statement according to which if a differential game has a continuously differentiable value function, then such function is a classical solution of the HJ equation associated to the game. This result strongly relies on Bellman\'s Dynamic Programming Principle - and this is the reason why we devote an entire chapter to this theme. Furthermore, HJ is still at our sight from the PDE point of view. Our motivation is simple: under some lack of regularity - a value function which is continuous, but not continuously differentiable - a game may still have a value function represented as a solution of the associated HJ equation. In this case such a solution will be called a solution in the viscosity sense. We then discuss the properties of viscosity solutions as well as provide an existence and uniqueness theorem. Finally we turn our attention back to the theory of games and - through the notion of viscosity - establish the existence and uniqueness of value functions for a differential game within viscosity solution theory.

Dans les installations nucléaires, les explosions, qu’elles soient d’origine interne ou externe, peuvent entrainer la rupture du confinement et le rejet de matières radioactives dans l’environnement. Il est donc fondamental, dans un cadre de sûreté de modéliser ce phénomène. L’objectif de cette thèse est de contribuer à l’élaboration de schémas numériques performants pour résoudre ces modèles complexes. Les travaux présentés s’articule autour de deux axes majeurs : le développement de schémas volumes finis consistants pour les équations d’Euler compressible qui modélise les ondes de choc et celui de schémas performants pour la propagation d’interfaces comme le front de flamme lors d'une déflagration. La discrétisation spatiale est de type mailles décalées pour tous les schémas développés. Les schémas pour les équations d'Euler se basent sur une formulation en énergie interne qui permet de préserver sa positivité ainsi que celle de la masse volumique. Un bilan d'énergie cinétique discret peut être obtenu et permet de retrouver un bilan d'énergie totale par l'ajout d'un terme de correction dans le bilan d'énergie interne. Le schéma ainsi construit est consistant au sens de Lax avec les solutions faibles entropiques des équations continues. On utilise les propriétés des équations de type Hamilton-Jacobi pour construire une classe de schémas volumes finis performants sur une large variété de maillages modélisant la propagation du front de flamme. Ces schémas garantissent un principe du maximum et possèdent des propriétés importantes de monotonie et consistance qui permettent d'obtenir un résultat de convergence. / In nuclear facilities, internal or external explosions can cause confinement breaches and radioactive materials release in the environment. Hence, modeling such phenomena is crucial for safety matters. The purpose of this thesis is to contribute to the creation of efficient numerical schemes to solve these complex models. The work presented here focuses on two major aspects: first, the development of consistent schemes for the Euler equations which model the blast waves, then the buildup of reliable schemes for the front propagation, like the flame front during the deflagration phenomenon. Staggered discretization is used in space for all the schemes. It is based on the internal energy formulation of the Euler system, which insures its positivity and the positivity of the density. A discrete kinetic energy balance is derived from the scheme and a source term is added in the discrete internal energy balance equation to preserve the exact total energy balance. High order, MUSCL-like interpolators are used in the discrete momentum operators. The resulting scheme is consistent (in the sense of Lax) with the weak entropic solutions of the continuous problem. We use the properties of Hamilton-Jacobi equations to build a class of finite volume schemes compatible with a large number of meshes to model the flame front propagation. These schemes satisfy a maximum principle and have important consistency and monotonicity properties. These latters allows to derive a convergence result for the schemes based on Cartesian grids.

[pt] A técnica de análise de sinais corrompidos por ruído baseada no comportamento de subespaços vetoriais foi tema de alguns trabalhos publicados desde o início da década de 80. Esta nova técnica passou a ter grande importância no processamento de sinais digitais devido a fatores como robustez e precisão.Porém, o maior problema associado a este novo método é o seu elevado custo computacional. Esta característica limitou o emprego da técnica em sistemas - offline - . A preocupação então passou a ser rastrear a variação do comportamento dos subespaços vetoriais de modo eficiente. O objetivo deste rastreamento seria o emprego da técnica em alguns sistemas que operam em tempo real. Este trabalho de tese propõe um novo algoritmo de rastreamento de subespaços vetoriais. O objetivo é apresentar um algoritmo que demonstre um bom desempenho, com relação aos demais já existentes, permitindo eventual aplicação em sistemas que atuem em tempo real. Como contribuição adicional, são apresentadas uma nova análise e caracterização de sistemas que se assemelham aos circulantes, sendo para isto reinterpretada a decomposição de matrizes circulantes. O conjunto de contribuições é aplicado a um novo sistema automático de classificação de sinais comunicação, quanto ao tipo de modulação. / [en] The signal subspace analysis technique, usually applied to signals corrupted by noise, is theme of some papers since the beginning of the 80s decade. This new technique has presented important features, as robustness and precision, and became widely employed in digital signal processing. However, the main problem associated to this new method is the high computational cost. This characteristic has restricted the use of signal subspace analysis to some off-line systems. A possible way to overcome this burden was to track the signal and noise subspace behavior in the time-domain. The main objective of these methods is to allow the signal subspace analysis technique application to real time systems, sometimes at the expense of limiting analysis precision or scope. This work proposes a new subspace tracking procedure. The goal is to describe a new algorithm with good performance (precision-speed), allowing some real time systems applications. A new analysis and characterization of almost circulant systems is introduced by reinterpreting the circulating matrix decomposition scheme. The set of contributions is applied to a new analogue modulation communication signals automatic recognition structure.

La Diagonalisation Conjointe par Congruence (DCC) d'un ensemble de matrices apparaît dans nombres de problèmes de traitement du signal, tels qu'en Analyse en Composantes Indépendantes (ACI). Les développements récents en ACI sous contrainte de non négativité de la matrice de mélange, nommée ACI semi-non négative, permettent de tirer profit d'une modélisation physique réaliste des phénomènes observés tels qu'en audio, en traitement d'image ou en ingénierie biomédicale. Par conséquent, durant cette thèse, l'objectif principal était non seulement de concevoir et développer des algorithmes d'ACI semi-non négative basés sur de nouvelles méthodes de DCC non négative où la matrice de passage recherchée est non négative, mais également d'illustrer leur intérêt dans le cadre d'applications pratiques de séparation de sources. Les algorithmes de DCC non négative proposés exploitent respectivement deux stratégies fondamentales d'optimisation. La première famille d'algorithmes comprend cinq méthodes semi-algébriques, reposant sur la méthode de Jacobi. Cette famille prend en compte la non négativité par un changement de variable carré, permettant ainsi de se ramener à un problème d'optimisation sans contrainte. L'idée générale de la méthode de Jacobi est de i) factoriser la matrice recherchée comme un produit de matrices élémentaires, chacune n'étant définie que par un seul paramètre, puis ii) d'estimer ces matrices élémentaires l'une après l'autre dans un ordre spécifique. La deuxième famille compte un seul algorithme, qui utilise la méthode des directions alternées. Un tel algorithme est obtenu en minimisant successivement le Lagrangien augmenté par rapport aux variables et aux multiplicateurs. Les résultats expérimentaux sur les matrices simulées montrent un gain en performance des algorithmes proposés par comparaison aux méthodes DCC classiques, qui n'exploitent pas la contrainte de non négativité. Il semble que nos méthodes peuvent fournir une meilleure précision d'estimation en particulier dans des contextes difficiles, par exemple, pour de faibles valeurs de rapport signal sur bruit, pour un petit nombre de matrices à diagonaliser et pour des niveaux élevés de cohérence de la matrice de passage. Nous avons ensuite montré l'intérêt de nos approches pour la résolution de problèmes pratiques de séparation aveugle de sources. Pour n'en citer que quelques-uns, nous sommes intéressés à i) l'analyse de composés chimiques en spectroscopie par résonance magnétique, ii) l'identification des profils spectraux des harmoniques (par exemple, de notes de piano) d'un morceau de musique mono-canal par décomposition du spectrogramme, iii) l'élimination partielle du texte se trouvant au verso d'une feuille de papier fin. Ces applications démontrent la validité et l'intérêt de nos algorithmes en comparaison des méthodes classique de séparation aveugle de source. / The Joint Diagonalization of a set of matrices by Congruence (JDC) appears in a number of signal processing problems, such as in Independent Component Analysis (ICA). Recent developments in ICA under the nonnegativity constraint of the mixing matrix, known as semi-nonnegative ICA, allow us to obtain a more realistic representation of some real-world phenomena, such as audios, images and biomedical signals. Consequently, during this thesis, the main objective was not only to design and develop semi-nonnegative ICA methods based on novel nonnegative JDC algorithms, but also to illustrate their interest in applications involving Blind Source Separation (BSS). The proposed nonnegative JDC algorithms belong to two fundamental strategies of optimization. The first family containing five algorithms is based on the Jacobi-like optimization. The nonnegativity constraint is imposed by means of a square change of variable, leading to an unconstrained problem. The general idea of the Jacobi-like optimization is to factorize the matrix variable as a product of a sequence of elementary matrices which is defined by only one parameter, then to estimate these elementary matrices one by one in a specific order. The second family containing one algorithm is based on the alternating direction method of multipliers. Such an algorithm is derived by successively minimizing the augmented Lagrangian function of the cost function with respect to the variables and the multipliers. Experimental results on simulated matrices show a better performance of the proposed algorithms in comparison with several classical JDC methods, which do not use the nonnegativity as constraint prior. It appears that our methods can achieve a better estimation accuracy particularly in difficult contexts, for example, for a low signal-to-noise ratio, a small number of input matrices and a high coherence level of matrix. Then we show the interest of our approaches in solving real-life problems. To name a few, we are interested in i) the analysis of the chemical compounds in the magnetic resonance spectroscopy, ii) the identification of the harmonically fixed spectral profiles (such as piano notes) of a piece of signal-channel music record by decomposing its spectrogram, iii) the partial removal of the show-through effect of digital images, where the show-through effect were caused by scanning a semi-transparent paper. These applications demonstrate the validity and improvement of our algorithms, comparing with several state-of-the-art BSS methods.

Dans cette thèse nous étudions une classe d’équations de Hamilton-Jacobi-Bellman provenant de la théorie du contrôle optimal des équations différentielles ordinaires. Nous nous intéressons principalement à l’analyse de la sensibilité de la fonction valeur des problèmes de contrôle optimal associés à de telles équations de H-J-B. Dans la littérature, les relations de sensibilité fournissent une “mesure” de la robustesse des stratégies optimales par rapport aux variations de la variable d’état. Ces résultats sont des outils très importants pour le contrôle appliqué, parce qu’ils permettent d’étudier les effets que des approximations des données du système peuvent avoir sur les politiques optimales. Cette thèse est dédiée également à l’étude des problèmes de Mayer et de temps minimal. Nous supposons que la dynamique du problème soit une inclusion différentielle, afin de permettre aux données d’être non régulières et d’embrasser un ensemble plus grand d’applications. Néanmoins, cette tâche rend notre analyse plus difficile. La première contribution de cette étude est une extension de quelques résultats classiques de la théorie de la sensibilité au domaine des problèmes non paramétrées. Ces relations prennent la forme d’inclusions d’état adjoint, figurant dans le principe du maximum de Pontryagin, dans certains gradients généralisés de la fonction valeur évalués le long des trajectoires optimales. En deuxième lieu, nous développons des nouvelles relations de sensibilité impliquant des approximations du deuxième ordre de la fonction valeur. Cette analyse mène à de nouvelles applications concernant la propagation, tant ponctuel que local, de la régularité de la fonction valeur le long des trajectoires optimales. Nous proposons également des applications aux conditions d’optimalité. / This dissertation investigates a class of Hamilton-Jacobi-Bellman equations arising in optimal control of O.D.E.. We mainly focus on the sensitivity analysis of the optimal value function associated with the underlying control problems. In the literature, sensitivity relations provide a measure of the robustness of optimal control strategies with respect to variations of the state variable. This is a central tool in applied control, since it allows to study the effects that approximations of the inputs of the system may produce on the optimal policies. In this thesis, we deal whit problems in the Mayer or in the minimum time form. We assume that the dynamic is described by a differential inclusion, in order to allow data to be nonsmooth and to embrace a large area of concrete applications. Nevertheless, this task makes our analysis more challenging. Our main contribution is twofold. We first extend some classical results on sensitivity analysis to the field of nonparameterized problems. These relations take the form of inclusions of the co-state, featuring in the Pontryagin maximum principle, into suitable gradients of the value function evaluated along optimal trajectories. Furthermore, we develop new second-order sensitivity relations involving suitable second order approximations of the optimal value function. Besides being of intrinsic interest, this analysis leads to new consequences regarding the propagation of both pointwise and local regularity of the optimal value functions along optimal trajectories. As applications, we also provide refined necessary optimality conditions for some class of differential inclusions.

Cette thèse se développe sur deux fronts: nous nous concentrons sur les méthodes numériques des problèmes de contrôle optimal, en particulier sur le Principe de la Programmation Dynamique et sur le Model Predictive Control (MPC) et nous présentons des applications de techniques de contrôle en biologie. Dans la première partie, nous considérons l'approximation d'un problème de contrôle optimal avec horizon infini, qui combine une première étape, basée sur MPC permettant d'obtenir rapidement une bonne approximation de la trajectoire optimal, et une seconde étape, dans la quelle l¿équation de Bellman est résolue dans un voisinage de la trajectoire de référence. De cette façon, on peux réduire une grande partie de la taille du domaine dans lequel on résout l¿équation de Bellman et diminuer la complexité du calcul. Le deuxième sujet est le contrôle des méthodes Level Set: on considère un problème de contrôle optimal, dans lequel la dynamique est donnée par la propagation d'un graphe à une dimension, contrôlé par la vitesse normale. Un état finale est fixé, l'objectif étant de le rejoindre en minimisant une fonction coût appropriée. On utilise la programmation dynamique grâce à une réduction d'ordre de l'équation utilisant la Proper Orthogonal Decomposition. La deuxième partie est dédiée à l'application des méthodes de contrôle en biologie. On présente un modèle décrit par une équation aux dérivées partielles qui modélise l'évolution d'une population de cellules tumorales. On analyse les caractéristiques du modèle et on formule et résout numériquement un problème de contrôle optimal concernant ce modèle, où le contrôle représente la quantité du médicament administrée. / This thesis is divided in two parts: in the first part we focus on numerical methods for optimal control problems, in particular on the Dynamic Programming Principle and on Model Predictive Control (MPC), in the second part we present some applications of the control techniques in biology. In the first part of the thesis, we consider the approximation of an optimal control problem with an infinite horizon, which combines a first step based on MPC, to obtain a fast but rough approximation of the optimal trajectory and a second step where we solve the Bellman equation in a neighborhood of the reference trajectory. In this way, we can reduce the size of the domain in which the Bellman equation can be solved and so the computational complexity is reduced as well. The second topic of this thesis is the control of the Level Set methods: we consider an optimal control, in which the dynamics is given by the propagation of a one dimensional graph, which is controlled by the normal velocity. A final state is fixed and the aim is to reach the trajectory chosen as a target minimizing an appropriate cost functional. To apply the Dynamic Programming approach we firstly reduce the size of the system using the Proper Orthogonal Decomposition. The second part of the thesis is devoted to the application of control methods in biology. We present a model described by a partial differential equation that models the evolution of a population of tumor cells. We analyze the mathematical and biological features of the model. Then we formulate an optimal control problem for this model and we solve it numerically.

Cette thèse est consacrée à l’étude de phénomènes de propagation et de concentration dans des modèles d’équations intégro-différentielles venant de la écologie. On étudie certaines équations de réaction-diffusion non locales apparaissant en dynamique de populations, ainsi que des modèles représentant l’évolution Darwinienne avec un mode de reproduction sexué.Dans une première partie, nous étudions la propagation spatiale pour une équation de réaction-diffusion ou la dispersion opère via un noyau de convolution à queue lourde. Nous mesurons de manière précise l’accélération du front de propagation de la solution. Nous proposons également une échelle adaptée pour mesurer les «petites» mutations. Dans les deux cas nous utilisons le formalisme des équations de Hamilton-Jacobi.Dans un second temps nous étudions un modèle de génétique quantitative, avec un mode de reproduction sexuée. Un petit paramètre mesure la déviation entre le trait des descendants est la moyenne des traits des parents. Dans le régime où ce paramètre est petit nous étudions l’existence de solutions stationnaires, puis le problème de Cauchy lié à ce modèle. Les solutions se concentrent autour des optima de sélection, sous la forme de perturbations de distributions Gaussiennes avec petite variance fixée par le paramètre. Notre analyse généralise le cas linéaire de la reproduction asexuée en utilisant des outils d’analyse perturbative. Enfin dans une dernière partie nous fournissons des simulations numériques et des méthodes mathématiques pour étudier la dynamique interne des équilibres dans le régime de petite variance, pour les deux modes de reproduction : asexué et sexué. / This manuscript tackles propagation and concentration phenomena in different integro-differential equations with a background in ecology. We study non local reaction-diffusion equations from population dynamics, and models for Darwinian evolution with a sexual or asexual mode of reproduction, with a preference for the former.In a first part, we study spatial propagation for a reaction diffusion equation where dispersion acts through a fat tailed kernel. We measure accurately the acceleration of the propagation front of the population. We propose as well a scaling well adapted to “small mutations” when we consider the model in the context of adaptative dynamics. This scaling is very natural following the previous spatial investigation. In both cases we look at the long time behavior and we use the Hamilton-Jacobi framework. Then we turn our attention towards a quantitative genetics model, with a sexual mode of reproduction, imposed by the “infinitesimal operator”. In this non-linear setting, a small parameter tunes the deviation between the phenotypic trait of the offspring and the mean of the traits of the parents. In the regime where this parameter is small, we prove existence of stationary solutions, and their local uniqueness. We also provide an example of non-uniqueness in the case where the selection function admits several extrema. We prove that the solution concentrates around the points of minimum of the selection function. The analysis is carried by the small perturbations of special profiles : Gaussian distributions with small variance fixed by the parameter.We then study the stability of the Cauchy problem associated to the previous model. This time we prove that at all times, for a well prepared initial data, the solutions is arbitrary close to a Gaussian distribution with small variance. The proof follows the framework of the previous : we use perturbative analysis tools, but this time an even more precise description of the correctors is needed and we linearize the equation to obtain it. In a final part we show numerical simulations and different mathematical approaches to study inside dynamics of phenotypic lineages in the regime of small variance, with a moving environement.

京都大学 / 新制・課程博士 / 博士(理学) / 甲第23673号 / 理博第4763号 / 新制||理||1683(附属図書館) / 京都大学大学院理学研究科数学・数理解析専攻 / (主査)准教授 市野 篤史, 教授 雪江 明彦, 教授 池田 保 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM