Seja $\\mathcal X=\\{\\mathcal X_t:\\, t\\geq0,\\, \\mathcal X_0=0\\}$ um passeio aleatório $\\beta$-estável em $\\mathbb Z$ com média zero e com taxas de saltos não-homogêneas $\\{\\tau_i^: i\\in\\mathbb Z\\}$, com $\\beta\\in(1,2]$ e $\\{\\tau_i: i\\in\\mathbb Z\\}$ sendo uma família de variáveis aleatórias independentes com distribuição marginal comum na bacia de atração de uma lei $\\alpha$-estável, com $\\alpha\\in(0,2]$. Nesta tese, obtemos resultados sobre o comportamento do processo $\\mathcal X_t$ para tempos longos, em particular, obtemos seu limite de escala. Quando $\\alpha\\in(0,1)$, o limite de escala é um processo $\\beta$-estável mudado de tempo pela inversa de um outro processo, o qual envolve o tempo local do processo $\\beta$-estável e um independente subordinador $\\alpha$-estável; chamamos o processo resultante de processo quase-estável. Para o caso $\\alpha\\in[1,2]$, o limite de escala é um ordinário processo $\\beta$-estável. Para $\\beta=2$ e $\\alpha\\in(0,1)$, o limite de escala é uma quase-difusão com medida de velocidade aleatória estudada por Fontes, Isopi e Newman (2002). Outros resultados sobre o comportamento de $\\mathcal X$ para tempos longos são envelhecimento e localização. Nós obtemos resultados de envelhecimento integrado e não-integrado para $\\mathcal X$ quando $\\alpha\\in(0,1)$. Relacionado à esses resultados, e possivelmente de interesse independente, consideramos o processo de armadilha definido por $\\{\\tau_{\\mathcal X_t}: t\\geq0\\}$, e obtemos seu limite de escala. Concluímos a tese com resultados sobre localização de $\\mathcal X$. Mostramos que ele pode ser localizado quando $\\alpha\\in(0,1)$, e que não pode ser localizado quando $\\alpha\\in(1,2]$, assim estendendo os resultados de Fontes, Isopi e Newman (1999) para o caso de passeios simples simétricos. / Let $\\mathcal X=\\{\\mathcal X_t:\\, t\\geq0,\\, \\mathcal X_0=0\\}$ be a mean zero $\\beta$-stable random walk on $\\mathbb Z$ with inhomogeneous jump rates $\\{\\tau_i^: i\\in\\mathbb Z\\}$, with $\\beta\\in(1,2]$ and $\\{\\tau_i: i\\in\\mathbb Z\\}$ is a family of independent random variables with common marginal distribution in the basin of attraction of an $\\alpha$-stable law with $\\alpha\\in(0,2]$. In this thesis we derive results about the long time behavior of this process, in particular its scaling limit. When $\\alpha\\in(0,1)$, the scaling limit is a $\\beta$-stable process time-changed by the inverse of another process, involving the local time of the $\\beta$-stable process and an independent $\\alpha$-stable subordinator; the resulting process may be called a quasistable process. For the case $\\alpha\\in[1,2]$, the scaling limit is an ordinary $\\beta$-stable process. For $\\beta=2$ and $\\alpha\\in(0,1)$, the scaling limit is a quasidiffusion with random speed measure studied by Fontes, Isopi and Newman (2002). Other results about the long time behavior of $\\mathcal X$ concern aging and localization. We obtain integrated and non integrated aging results for $\\mathcal X$ when $\\alpha\\in(0,1)$. Related to these results, and possibly of independent interest, we consider the trap process defined as $\\{\\tau_{\\mathcal X_t}: t\\geq0\\}$, and derive its scaling limit. We conclude the thesis with results about localization of $\\mathcal X$. We show that it localizes when $\\alpha\\in(0,1)$, and does not localize when $\\alpha\\in(1,2]$, extending results of Fontes, Isopi and Newman (1999) for the simple symmetric case.

Seja $\\mathcal X=\\{\\mathcal X_t:\\, t\\geq0,\\, \\mathcal X_0=0\\}$ um passeio aleatório $\\beta$-estável em $\\mathbb Z$ com média zero e com taxas de saltos não-homogêneas $\\{\\tau_i^: i\\in\\mathbb Z\\}$, com $\\beta\\in(1,2]$ e $\\{\\tau_i: i\\in\\mathbb Z\\}$ sendo uma família de variáveis aleatórias independentes com distribuição marginal comum na bacia de atração de uma lei $\\alpha$-estável, com $\\alpha\\in(0,2]$. Nesta tese, obtemos resultados sobre o comportamento do processo $\\mathcal X_t$ para tempos longos, em particular, obtemos seu limite de escala. Quando $\\alpha\\in(0,1)$, o limite de escala é um processo $\\beta$-estável mudado de tempo pela inversa de um outro processo, o qual envolve o tempo local do processo $\\beta$-estável e um independente subordinador $\\alpha$-estável; chamamos o processo resultante de processo quase-estável. Para o caso $\\alpha\\in[1,2]$, o limite de escala é um ordinário processo $\\beta$-estável. Para $\\beta=2$ e $\\alpha\\in(0,1)$, o limite de escala é uma quase-difusão com medida de velocidade aleatória estudada por Fontes, Isopi e Newman (2002). Outros resultados sobre o comportamento de $\\mathcal X$ para tempos longos são envelhecimento e localização. Nós obtemos resultados de envelhecimento integrado e não-integrado para $\\mathcal X$ quando $\\alpha\\in(0,1)$. Relacionado à esses resultados, e possivelmente de interesse independente, consideramos o processo de armadilha definido por $\\{\\tau_{\\mathcal X_t}: t\\geq0\\}$, e obtemos seu limite de escala. Concluímos a tese com resultados sobre localização de $\\mathcal X$. Mostramos que ele pode ser localizado quando $\\alpha\\in(0,1)$, e que não pode ser localizado quando $\\alpha\\in(1,2]$, assim estendendo os resultados de Fontes, Isopi e Newman (1999) para o caso de passeios simples simétricos. / Let $\\mathcal X=\\{\\mathcal X_t:\\, t\\geq0,\\, \\mathcal X_0=0\\}$ be a mean zero $\\beta$-stable random walk on $\\mathbb Z$ with inhomogeneous jump rates $\\{\\tau_i^: i\\in\\mathbb Z\\}$, with $\\beta\\in(1,2]$ and $\\{\\tau_i: i\\in\\mathbb Z\\}$ is a family of independent random variables with common marginal distribution in the basin of attraction of an $\\alpha$-stable law with $\\alpha\\in(0,2]$. In this thesis we derive results about the long time behavior of this process, in particular its scaling limit. When $\\alpha\\in(0,1)$, the scaling limit is a $\\beta$-stable process time-changed by the inverse of another process, involving the local time of the $\\beta$-stable process and an independent $\\alpha$-stable subordinator; the resulting process may be called a quasistable process. For the case $\\alpha\\in[1,2]$, the scaling limit is an ordinary $\\beta$-stable process. For $\\beta=2$ and $\\alpha\\in(0,1)$, the scaling limit is a quasidiffusion with random speed measure studied by Fontes, Isopi and Newman (2002). Other results about the long time behavior of $\\mathcal X$ concern aging and localization. We obtain integrated and non integrated aging results for $\\mathcal X$ when $\\alpha\\in(0,1)$. Related to these results, and possibly of independent interest, we consider the trap process defined as $\\{\\tau_{\\mathcal X_t}: t\\geq0\\}$, and derive its scaling limit. We conclude the thesis with results about localization of $\\mathcal X$. We show that it localizes when $\\alpha\\in(0,1)$, and does not localize when $\\alpha\\in(1,2]$, extending results of Fontes, Isopi and Newman (1999) for the simple symmetric case.

Im ersten Teil der Dissertation wird ein diskretes stochastisches zustandsabhängiges Modell eines zweiseitigen Limit Orderbuchs als bestehend aus den Zustandsgrößen bester Bidpreis (Geldkurs), bester Askpreis (Briefkurs) und vorhandener Kauf- bzw. Verkaufsdichte definiert. Für eine einfache Skalierung mit zwei Zeitskalen wird ein Grenzwertsatz bewiesen. Die Veränderungen der besten Bid- und Askpreise werden im Sinne des Gesetzes der großen Zahlen skaliert und dies entspricht der langsameren Zeitskala. Das Platzieren bzw. Stornieren der Limitorder findet auf der schnelleren Zeitskala statt. Der Grenzwertsatz besagt, dass die fundamentalen Zustandsgrößen, gegeben Regularitätsbedingungen der einkommenden Order, fast sicher zu einem stetigen Limesmodell konvergieren. Im Limesmodell sind der beste Bidpreis und der beste Askpreis die eindeutigen Lösungen von zwei gekoppelten gewöhnlichen DGLen. Die Kauf- und Verkaufsdichten sind jeweils als eindeutige Lösungen von linearen hyperbolischen PDGLen, die anhand der Erwartungswerte der einkommenden Orderparameter festgelegt sind, gegeben. Die Lösungen sind in geschlossener Form erhältlich. Im zweiten Teil wird ein funktionaler zentraler Grenzwertsatz d.h. ein Invarianzprinzip für ein vereinfachtes Modell eines Limitorderbuches bewiesen. Unter einer natürlichen Skalierung konvergiert der zweidimensionale Preisprozess (Bid- und Askpreis) in Verteilung zu einer Semimartingal reflektierten Brownschen Bewegung in der zugelassenen Preismenge. Gleichzeitig konvergieren die Kauf- und Verkaufsdichten im schwachen Sinn zum Betrag einer zweiparametrischen Brownschen Bewegung. Es wird weiterhin anhand eines Beispiels gezeigt, wie man für das Modell im ersten Teil eine stochastiche PDGL, unter einer starken Stationaritätsannahme für die Orderplatzierungen und -stornierungen, herleiten kann. Im dritten Teil wird ein Mittelungs- bzw. ein Invarianzprinzip für diskrete Banach- bzw. Hilbertraumwertige stochastische Prozesse bewiesen. / In the first part of the thesis, we define a random state-dependent discrete model of a two-sided limit order book in terms of its key quantities best bid [ask] price and the standing buy [sell] volume density. For a simple scaling that introduces a slow time scaling, that is equivalent to the classical law of large numbers, for the bid/ask prices and a faster time scale for the limit volume placements/cancelations, that keeps the expected volume rate over the considered price interval invariant, we prove a limit theorem. The limit theorem states that, given regularity conditions on the random order flow, the key quantities converge in the sense of a strong law of large numbers to a tractable continuous limiting model. The limiting model is such that the best bid and ask price dynamics can be described in terms of two coupled ODE:s, while the dynamics of the relative buy and sell volume density functions are given as the unique solutions of two linear first-order hyperbolic PDE:s with variable coefficients, specified by the expectation of the order flow parameters. In the second part, we prove a functional central limit theorem i.e. an invariance principle for an order book model with block shaped volume densities close to the spread. The weak limit of the two-dimensional price process (best bid and ask price) is given by a semi-martingale reflecting Brownian motion in the set of admissible prices. Simultaneously, the relative buy and sell volume densities close to the spread converge weakly to the modulus of a two-parameter Brownian motion. We also demonstrate an example how to easily derive an SPDE for the relative volume densities in a simple case, when a strong stationarity assumption is made on the limit order placements and cancelations for the model suggested in the first part. In the third and final part of the thesis, we prove an averaging and an invariance principle for discrete processes taking values in Banach and Hilbert spaces, respectively.

Cette thèse se divise en deux parties. Nous nous intéressons dans un premier temps à des fragmentations d'arbres aléatoires, et aux arbres des coupes associés. Dans le cadre discret, les modèles étudiés sont des arbres de Galton-Watson, fragmentés en enlevant successivement des arêtes choisies au hasard. Nous étudions également leurs analogues continus, l'arbre brownien et les arbres stables, que l'on fragmente en supprimant des points donnés par des processus ponctuels de Poisson. L'arbre des coupes associé à l'un de ces processus, discret ou continu, décrit la généalogie des composantes connexes créées au fur et à mesure de la dislocation. Pour une fragmentation qui se concentre autour de nœuds de grand degré, nous montrons que l'arbre des coupes continu est la limite d'échelle des arbres des coupes discrets correspondants. Dans les cas brownien et stable, nous montrons également que l'on peut reconstruire l'arbre initial à partir de son arbre des coupes et d'un étiquetage bien choisi de ses points de branchement. Nous étudions ensuite un problème portant sur les cartes aléatoires, et plus précisément sur la quadrangulation uniforme infinie du plan (UIPQ). De récents résultats montrent que dans l'UIPQ, toutes les géodésiques infinies issues de la racine sont essentiellement similaires. Nous déterminons la quadrangulation limite obtenue en ré-enracinant l'UIPQ ''à l'infini'' sur de l'une de ces géodésiques. Cette étude se fait en découpant l'UIPQ le long de cette géodésique. Nous étudions les deux parties ainsi créées via une correspondance avec des arbres discrets, puis nous obtenons la limite souhaitée par recollement. / This PhD thesis is divided into two parts. First, we study some fragmentations of random trees and the associated cut-trees. The discrete models we are interested in are Galton-Watson trees, which are cut down by recursively removing random edges. We also consider their continuous counterparts, the Brownian and stable trees, which are fragmented by deleting the atoms of Poisson point processes. For these discrete and continuous models, the associated cut-tree describes the genealogy of the connected components which appear during the cutting procedure. We show that for a ''vertex-fragmentation'', in which the nodes having a large degree are more susceptible to be deleted, the continuous cut-tree is the scaling limit of the corresponding discrete cut-trees. In the Brownian and stable cases, we also give a transformation which rebuilds the initial tree from its cut-tree and a well chosen labeling of its branchpoints. The second part relates to random maps, and more precisely the uniform infinite quadrangulation of the plane (UIPQ). Recent results show that in the UIPQ, all infinite geodesic rays originating from the root are essentially similar. We identify the limit quadrangulation obtained by rerooting the UIPQ at a point ''at infinity'' on one of these geodesics. To do this, we split the UIPQ along this geodesic ray. Using a correspondence with discrete trees, we study the two sides, and obtain the desired limit by gluing them back together.