We present results that span three interconnected domains. Initially, our analysis is centred on Backward Stochastic Differential Equations (BSDEs) featuring time-delayed generators. Subsequently, we direct our interest towards Mean Field Games (MFGs) incorporating absorption aspects, with a focus on the corresponding Master Equation within a confined domain under the imposition of Dirichlet boundary conditions. The investigation culminates in exploring pertinent Machine Learning methodologies applied to financial and economic decision-making processes.

Diese Arbeit behandelt zwei Gebiete: stochastische partielle Rückwerts-Differentialgleichungen (BSPDEs) und Mean-Field-Games (MFGs). Im ersten Teil wird über eine stochastische Variante der De Giorgischen Iteration ein Maximumprinzip für quasilineare reflektierte BSPDEs (RBSPDEs) auf allgemeinen Gebieten bewiesen. Als Folgerung erhalten wir ein Maximumprinzip für RBSPDEs auf beschränkten, sowie für BSPDEs auf allgemeinen Gebieten. Abschließend wird das lokale Verhalten schwacher Lösungen untersucht. Im zweiten Teil zeigen wir zunächst die Existenz von Gleichgewichten in MFGs mit singulärer Kontrolle. Wir beweisen, dass die Lösung eines MFG ohne Endkosten und ohne Kosten in der singulären Kontrolle durch die Lösungen eines MFGs mit strikt regulären Kontrollen approximiert werden kann. Die vorgelegten Existenz- und Approximationsresultat basieren entscheidend auf der Wahl der Storokhod M1 Topologie auf dem Raum der Càdlàg-Funktion. Anschließend betrachten wir ein MFG optimaler Portfolioliquidierung unter asymmetrischer Information. Die Lösung des MFG charakterisieren wir über eine stochastische Vorwärts-Rückwärts-Differentialgleichung (FBSDE) mit singulärer Endbedingung der Rückwärtsgleichung oder alternativ über eine FBSDE mit endlicher Endbedingung, jedoch singulärem Treiber. Wir geben ein Fixpunktargument, um die Existenz und Eindeutigkeit einer Kurzzeitlösung in einem gewichteten Funktionenraum zu zeigen. Dies ermöglicht es, das ursprüngliche MFG mit entsprechenden MFGs ohne Zustandsendbedinung zu approximieren. Der zweite Teil wird abgeschlossen mit einem Leader-Follower-MFG mit Zustandsendbedingung im Kontext optimaler Portfolioliquidierung bei hierarchischer Agentenstruktur. Wir zeigen, dass das Problem beider Spielertypen auf singuläre FBSDEs zurückgeführt werden kann, welche mit ähnlichen Methoden wie im vorangegangen Abschnitt behandelt werden können. / The thesis is concerned with two topics: backward stochastic partial differential equations and mean filed games. In the first part, we establish a maximum principle for quasi-linear reflected backward stochastic partial differential equations (RBSPDEs) on a general domain by using a stochastic version of De Giorgi’s iteration. The maximum principle for RBSPDEs on a bounded domain and the maximum principle for BSPDEs on a general domain are obtained as byproducts. Finally, the local behavior of the weak solutions is considered. In the second part, we first establish the existence of equilibria to mean field games (MFGs) with singular controls. We also prove that the solutions to MFGs with no terminal cost and no cost from singular controls can be approximated by the solutions, respectively control rules, for MFGs with purely regular controls. Our existence and approximation results strongly hinge on the use of the Skorokhod M1 topology on the space of càdlàg functions. Subsequently, we consider an MFG of optimal portfolio liquidation under asymmetric information. We prove that the solution to the MFG can be characterized in terms of a forward backward stochastic differential equation (FBSDE) with possibly singular terminal condition on the backward component or, equivalently, in terms of an FBSDE with finite terminal value, yet singular driver. We apply the fixed point argument to prove the existence and uniqueness on a short time horizon in a weighted space. Our existence and uniqueness result allows to prove that our MFG can be approximated by a sequence of MFGs without state constraint. The final result of the second part is a leader follower MFG with terminal constraint arising from optimal portfolio liquidation between hierarchical agents. We show the problems for both follower and leader reduce to the solvability of singular FBSDEs, which can be solved by a modified approach of the previous result.

Ce manuscrit présente les fondements dynamiques des jeux de population avec un nombre variable de joueurs ainsi que leurs concepts de solutions et de stabilités. Nous introduisons d'abord les dynamiques de jeux avec retard et étudions leurs stabilités. Nous les appliquons aux réseaux filaires et aux réseaux sans fils. Ensuite nous nous intéressons aux aspects de mobilité et aux distributions spatiales des joueurs sur le réseau. Cela nous conduit à une nouvelle classe de dynamique de jeux à stratégies vectorielles avec des contraintes de migrations, appelée dynamique de jeux d'évolution avec migration. Nous dérivons de telles dynamiques pour les réseaux hybrides et appliquons aux problèmes de contrôle de puissance dans les réseaux hétérogènes, choix entre plusieurs technologies et migration entre plusieurs classes d'utilisateurs. Ensuite nous nous focalisons aux jeux stochastiques de population avec plusieurs classes de joueurs dans lesquels chaque joueur possède son propre état et fait face un vecteur qui évolue dans le temps. Des applications à la gestion d'énergie dans les réseaux sont présentées. Finalement, nous étudions une classe de jeux à champ moyen. Lorsque la taille de la population devient très grande, les asymptotiques du système conduisent à des dynamiques appelées dynamiques de jeux à champ moyen. Cette classe de dynamiques contient les dynamiques standard basées sur des révisions de stratégies. Nous utilisons ce modèle pour analyser les problèmes accès aléatoires à des ressources dans un environnement où les utilisateurs et les ressources sont spatialement distribuées. Nous établissons un lien entre les jeux à champ moyen et les jeux différentiels de population dans lesquels chaque joueur a son état individuel et optimise son paiement à long terme pendant son temps de séjour dans le système sous contraintes que le profil de population évolue selon une dynamique de jeux à champ moyen / His manuscript presents dynamic foundations of population games with variable number of players and their solutions and stability concepts. We first introduce delayed evolutionary game dynamics and study their stability. Applications to both wired and wireless networks are presented. We then introduce mobility and spatial aspects of players distribution into the network dynamics. This leads to a new class of game dynamics with multicomponent strategies and migration constraints called evolutionary game dynamics with migration. We derived such dynamics for hybrid systems such as power control in heterogenous networks, switching between technologies and migration between different classes of users. After that we focus on stochastic population games with multiple classes of players in which each player has its own state and facing to an evolving vector which represents the population profile. We use this model to analyze resource and energy constrained interactions in wireless networks. Finally, we present a class of mean field games. When taking the asymptotics of finite systems, we derive a new class of game dynamics called mean field game dynamics. This class contains the standard evolutionary game dynamics based on revision of pure actions. We apply this model to analyze spatial random access game and dynamic resource competition game with individual states. We establish a link betweenmean field games and differential population games inwhich each player optimizes its long-term objective during its sojourn time in the system subject to the constraint that the population profile evolves according to some mean field game dynamics

Submitted by Aelson Maciera (aelsoncm@terra.com.br) on 2017-06-27T20:42:50Z No. of bitstreams: 1 DissLFL.pdf: 818058 bytes, checksum: a45e8f4dbdc692c6f31fde1d45f6574d (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-07-03T17:56:46Z (GMT) No. of bitstreams: 1 DissLFL.pdf: 818058 bytes, checksum: a45e8f4dbdc692c6f31fde1d45f6574d (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-07-03T17:56:52Z (GMT) No. of bitstreams: 1 DissLFL.pdf: 818058 bytes, checksum: a45e8f4dbdc692c6f31fde1d45f6574d (MD5) / Made available in DSpace on 2017-07-03T18:01:59Z (GMT). No. of bitstreams: 1 DissLFL.pdf: 818058 bytes, checksum: a45e8f4dbdc692c6f31fde1d45f6574d (MD5) Previous issue date: 2016-12-20 / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / In this thesis, we present a priori estimates for solutions of a mean-field game (MFG) defined over a bounded domain Ω ⊂ ℝd. We propose an application of these results to a model of capital and wealth accumulation. In Chapter 1, an introduction to mean-field games is presented. We also put forward some of the motivation from Economics and discuss previous developments in the theory of differential games. These comments aim at indicating the connection between mean-field games theory, its applications and the realm of Mathematical Analysis. In Chapter 2, we present an optimal control problem. Here, the agents are supposed to be undistinguishable, rational and intelligent. Undistinguishable means that every agent is governed by the same stochastic differential equation. Rational means that all efforts of the agent is to maximize a payoff functional. Intelligent means that they are able to solve an optimal control problem. Once we describe this (stochastic) optimal control problem, we produce a heuristic derivation of the mean-field games system, which is summarized in a Verification Theorem; this gives rise to the Hamilton-Jacobi equation (HJ). After that, we obtain the Fokker-Plank equation (FP). Finally, we present a representation formula for the solutions to the (HJ) equation, together with some regularity results. In Chapter 3, a specific optimal control problem is described and the associated MFG is presented. This MFG is prescribed in a bounded domain Ω ⊂ ℝd, which introduces substantialadditional challenges from the mathematical view point. This is due to estimates for the solutionsat the boundary in Lp. The rest of the chapter puts forward two well known tips of estimates: theso-called Hopf-Lax formula and the First Order Estimate. In Chapter 4, the wealth and capital accumulation mean-field game model is presented. The relevance of studying MFG in a bounded domain then becomes clear. In light of the results obtained in Chapter 3, we close Chapter 4 with the Hopf-Lax formula, and the First Order estimates. Three appendices close this thesis. They gather elementary material on Stochastic Calculus and Functional Analysis. / Nesta dissertação são apresentadas algumas estimativas a priori para soluções de sistemas mean-field games (MFG), definidos em domínios limitados Ω ⊂ ℝd. Tais estimativas são aplicadas em um modelo mean-field específico, que descreve o acúmulo de riqueza e capital. No Capítulo 1, é apresentada uma breve introdução histórica sobre os mean-field games. Nesta introdução, exploramos sua relação com a teoria dos jogos, cujos alicerces foram construídos por economistas e matemáticos ao longo do século XX. O objetivo do capítulo é transmitir. No Capítulo 2, apresentamos um problema de controle ótimo em que cada agente é suposto ser indistinguível, racional e inteligente. Indistinguível no sentido de que cada um é governado pela mesma equação diferencial estocástica. Racional no sentido de que todos os esforços do agente são no sentido de maximizar um funcional de recompensa e, inteligente no sentido de que são capazes de resolver um problema de controle ótimo. Descreve-se este problema de controle ótimo, e apresenta-se a derivação heurística dos mean-field games; obtém-se através de um Teorema de Verificação, a equação de Hamilton-Jacobi (HJ) associada, e em seguida, obtémse a equação de Fokker-Planck. De posse destas equações, apresentamos alguns resultados preliminares, como uma fórmula de representação para soluções da equação de HJ e alguns resultados de regularidade. No Capítulo 3, descreve-se um problema específico de controle ótimo e apresenta-se a respectiva derivação heurística culminando na descrição de um MFG com condições não periódicas na fronteira; esta abordagem é original na literatura de MFG. O restante do capítulo é dedicado à exposição de dois tipos bem conhecidos de estimativas: a fórmula de Hopf-Lax e estimativa de Primeira Ordem. Uma observação relevante, é a de que o trabalho em obter-se estimativas a priori é aumentado substancialmente neste caso, devido ao fato de lidarmos com estimativas para os termos de fronteira com normas em Lp. ao leitor, as origens da Teoria Econômica contemporânea, que surgem à partir da utilização da Matemática na formulação e resolução de problemas econômicos. Tal abordagem é motivada principalmente pelo rigor e clareza da Matemática em tais circunstâncias. No Capítulo 4, apresenta-se o modelo de jogo do tipo mean-field de acúmulo de capital e riqueza, o que deixa claro a relevância do estudo dos MFG em um domínio limitado. À luz dos resultados obtidos no Capítulo 3, encerramos o Capítulo 4 com as estimativas do tipo Hopf-Lax e de Primeira Ordem. Três apêndices encerram o texto desta dissertação de mestrado; estes reúnem material elementar sobre Cálculo Estocástico e Análise Funcional.

[pt] Neste trabalho examinamos equações completamente não-lineares em dois contextos distintos. A princípio, estudamos jogos de campo médio completamente não-lineares. Aqui, examinamos ganhos de regularidade para as soluções do problema, existência de soluções, resultados de relaxação e aspectos particulares de um example explícito. A segunda metade da tese dedica-se à regularidade ótima das soluções de um modelo completamente não-linear que degenera-se com respeito ao gradiente das soluções. A pergunta fundamental subjacente a ambos os tópicos diz respeito aos efeitos da elipticidade sobre propriedades intrínsecas das soluções de equações não-lineares. Mais precisamente, no caso dos jogos de campo médio, a elipticidade parece magnificada pelos efeitos do acoplamento, enquanto no caso dos problemas degenerados, esta quantidade colapsa em sub-regiões do domínio, dando origem a delicados fenômenos. Nossa análise inclui um breve contexto da inserção do trabalho. / [en] In this thesis, we examine fully nonlinear problems in two distinct contexts. The first part of our work focuses on fully nonlinear mean-field games. In this context, we examine gains of regularity, the existence of solutions, relaxation results, and particular aspects of a one-dimensional problem. The second half of the thesis concerns a (sharp) regularity theory for fully nonlinear equations degenerating with respect to the gradient of the solutions. The fundamental question underlying both topics regards the effects of ellipticity on the intrinsic properties of solutions to nonlinear equations. To be more precise, in the case of mean-field game systems, ellipticity seems to be magnified through the coupling structure. On the other hand, in the degenerate setting, ellipticity collapses, giving rise to intricate regularity phenomena. Our analysis is preceded by some context on both topics.

Cette thèse propose une étude mathématique des stratégies de vaccination.La partie I présente le cadre mathématique, notamment le modèle à compartiments Susceptible - Infected – Recovered.La partie II aborde les techniques mathématiques de type contrôle optimal employées afin de trouver une stratégie optimale de vaccination au niveau de la société. Ceci se fait en minimisant le coût de la société. Nous montrons que la fonction valeur associée peut avoir une régularité plus faible que celle attendue dans la littérature. Enfin, nous appliquons les résultats à la vaccination contre la coqueluche.La partie III présente un modèle où le coût est défini au niveau de l'individu. Nous reformulons le problème comme un équilibre de Nash et comparons le coût obtenu avec celui de la stratégie sociétale. Une application à la grippe A(H1N1) indique la présence de perceptions différentes liées à la vaccination.La partie IV propose une implémentation numérique directe des stratégies présentées. / This thesis propose a mathematical analysis of the vaccination strategies.The first part introduces the mathematical framework, in particular the Susceptible – Infected – Recovered compartmental model.The second part introduces the optimal control tools used to find an optimal vaccination strategy from the societal point of view, which is a minimizer of the societal cost. We show that the associated value function can have a less regularity than what was assumed in the literature. These results are then applied to the vaccination against the whooping cough.The third part defines a model where the cost is defined at the level of the individual. We rephrase this problem as a Nash equilibrium and compare this results with the societal strategy. An application to the Influenza A(H1N1) 2009-10 indicates the presence of inhomogeneous perceptions concerning the vaccination risks.The fourth and last part proposes a direct numerical implementation of the different strategies.