The thesis covers methods for image segmentation. Fuzzy segmentation is based on the thresholding method. This is generalized to accept multiple criteria. The whole process is mathematically based on the free algebra theory. Free distributive lattice is created from poset of elements based on image properties and the lattice members are represented by terms used by the threshoding. Possible segmentation results compose the equivalence classes distribution. The thesis also contains description of resulting algorithms and methods for their optimization. Also the method of area subtracting is introduced.

Traditionnellement, les applications orientées objets légataires intègrent différents aspects fonctionnels. Ces aspects peuvent être dispersés partout dans le code. Il existe différents types d’aspects : • des aspects qui représentent des fonctionnalités métiers ; • des aspects qui répondent à des exigences non fonctionnelles ou à d’autres considérations de conception comme la robustesse, la distribution, la sécurité, etc. Généralement, le code qui représente ces aspects chevauche plusieurs hiérarchies de classes. Plusieurs chercheurs se sont intéressés à la problématique de la modularisation de ces aspects dans le code : programmation orientée sujets, programmation orientée aspects et programmation orientée vues. Toutes ces méthodes proposent des techniques et des outils pour concevoir des applications orientées objets sous forme de composition de fragments de code qui répondent à différents aspects. La séparation des aspects dans le code a des avantages au niveau de la réutilisation et de la maintenance. Ainsi, il est important d’identifier et de localiser ces aspects dans du code légataire orienté objets. Nous nous intéressons particulièrement aux aspects fonctionnels. En supposant que le code qui répond à un aspect fonctionnel ou fonctionnalité exhibe une certaine cohésion fonctionnelle (dépendances entre les éléments), nous proposons d’identifier de telles fonctionnalités à partir du code. L’idée est d’identifier, en l’absence des paradigmes de la programmation par aspects, les techniques qui permettent l’implémentation des différents aspects fonctionnels dans un code objet. Notre approche consiste à : • identifier les techniques utilisées par les développeurs pour intégrer une fonctionnalité en l’absence des techniques orientées aspects • caractériser l’empreinte de ces techniques sur le code • et développer des outils pour identifier ces empreintes. Ainsi, nous présentons deux approches pour l’identification des fonctionnalités existantes dans du code orienté objets. La première identifie différents patrons de conception qui permettent l’intégration de ces fonctionnalités dans le code. La deuxième utilise l’analyse formelle de concepts pour identifier les fonctionnalités récurrentes dans le code. Nous expérimentons nos deux approches sur des systèmes libres orientés objets pour identifier les différentes fonctionnalités dans le code. Les résultats obtenus montrent l’efficacité de nos approches pour identifier les différentes fonctionnalités dans du code légataire orienté objets et permettent de suggérer des cas de refactorisation. / Object oriented applications integrate various functional aspects. These aspects can be scattered everywhere in the code. There are various types of aspects : • aspects which represent business functionalities ; • aspects related to non functional requirements or to design concerns such as robustness, distribution, and security. The code representing such aspects can be located in different class hierarchies. Researchers have been interested in the problem of the modularisation of these aspects and many approaches were proposed : oriented programming subjects, oriented programming Aspects and oriented programming view. These approaches offer techniques and tools for designing object oriented applications based on the composition of slices of various aspects. The main benefit of the separation of aspects is supporting reuse and maintenance. Consequently, it is well worth identifying and extracting aspects of legacy object oriented applications. Our work mainly focuses on functional aspects. Assuming that the code of a functional aspect or a feature has a functional cohesion (dependencies between elements), we suggest methods for identifying such features from the code. The idea is to identify, in the absence of any aspect oriented paradigm, the techniques used for implementing a feature in the code. Our approach consists of : • identifying techniques used by developers to integrate a feature in the absence of aspect oriented techniques • characterizing the patterns of these techniques • and developing tools to identify these patterns. We present two approaches for the identification of the existing features in the object oriented code. The first one identifies various design patterns which integrates these features in the code. The second approach uses the formal concept analysis to identify the recurring features in the code. We experiment our approaches to identify functional features in different open source object oriented applications. The results show the efficiency of our approaches in identifying various functional features in the legacy object oriented, and can some times suggest refactoring.

"Fluid-structure interaction arises in a wide array of technological applications including naval and marine hydrodynamics, civil and wind engineering and flight vehicle aerodynamics. When a fluid flows over a bluff body such as a circular cylinder, the periodic vortex shedding in the wake causes fluctuating lift and drag forces on the body. This phenomenon can lead to fatigue damage of the structure due to large amplitude vibration. It is widely believed that the wake structures behind the structure determine the hydrodynamic forces acting on the structure and control of wake structures can lead to vibration control of the structure. Modeling this complex non-linear interaction requires coupling of the dynamics of the fluid and the structure. In this thesis, however, the vibration of the flexible cylinder is prescribed, and the focus is on modeling the fluid dynamics in its wake. Low-dimensional iterative circle maps have been found to predict the universal dynamics of a two-oscillator system such as the rigid cylinder wake. Coupled map lattice (CML)models that combine a series of low-dimensional circle maps with a diffusion model have previously predicted qualitative features of wake patterns behind freely vibrating cables at low Reynolds number. However, the simple nature of the CML models implies that there will always be unmodelled wake dynamics if a detailed, quantitative comparison is made with laboratory or simulated wake flows. Motivated by a desire to develop an improved CML model, we incorporate self-learning features into a new CML that is trained to precisely estimate wake patterns from target numerical simulations and experimental wake flows. The eventual goal is to have the CML learn from a laboratory flow in real time. A real-time self-learning CML capable of estimating experimental wake patterns could serve as a wake model in a future anticipated feedback control system designed to produce desired wake patterns. A new convective-diffusive map that includes additional wake dynamics is developed. Two different self-learning CML models, each capable of precisely estimating complex wake patterns, have been developed by considering additional dynamics from the convective-diffusive map. The new self-learning CML models use adaptive estimation schemes which seek to precisely estimate target wake patterns from numerical simulations and experiments. In the first self-learning CML, the estimator scheme uses a multi-variable least-squares algorithm to adaptively vary the spanwise velocity distribution in order to minimize the state error (difference between modeled and target wake patterns). The second self-learning model uses radial basis function neural networks as online approximators of the unmodelled dynamics. Additional unmodelled dynamics not present in the first self-learning CML model are considered here. The estimator model uses a combination of a multi-variable normalized least squares scheme and a projection algorithm to adaptively vary the neural network weights. Studies of this approach are conducted using wake patterns from spectral element based NEKTAR simulations of freely vibrating cable wakes at low Reynolds numbers on the order of 100. It is shown that the self-learning models accurately and efficiently estimate the simulated wake patterns within several shedding cycles. Next, experimental wake patterns behind different configurations of rigid cylinders were obtained. The self-learning CML models were then used for off-line estimation of the stored wake patterns. With the eventual goal of incorporating low-order CML models into a wake pattern control system in mind, in a related study control terms were added to the simple CML model in order to drive the wake to the desired target pattern of shedding. Proportional, adaptive proportional and non-linear control techniques were developed and their control efficiencies compared."

Cette thèse décrit l'observation expérimentale de l'effet Hong-Ou-Mandel avec une sourceatomique ultra-froide. L’expérience originale réalisée en 1987 par C. K. Hong, Z. Y. Ou et L. Mandel illustre de façon simple une interférence à deux particules explicable uniquement par la mécanique quantique : deux particules bosoniques et indiscernables, arrivant chacune sur une face d'entrée différente d'une lame semi-réfléchissante ressortent ensemble. Cet effet se traduit par une réduction du taux de détection en coïncidence entre les deux voies de sortie quand les particules arrivent simultanément sur la lame. Cette expérience fut originalement réalisée avec des photons et nous rapportons ici la première mise en oeuvre expérimentale avec des particules massives se propageant dans l’espace libre.Après présentation des différentes techniques nécessaires à sa réalisation, nous décrivons cette expérience et analysons les résultats obtenus. En particulier, la réduction du taux de coïncidence est suffisamment forte pour exclure toute interprétation classique ; l'observation de cet effet constitue une brique fondamentale dans le domaine de l’information quantique atomique. / In this thesis, we report the first realisation of the Hong–Ou–Mandel experiment with massive particles in momentum space. This milestone experiment was originally performed in quantum optics: two photons arriving simultaneously at the input ports of a 50:50 beam-splitter always emerge together in one of the output ports. The effect leads to a reduction of coincidence counts which translates into a dip when particles are indistinguishable. We performed the experiment with metastable helium atoms where the specificities of the Micro-Channel-Plate detector allows one to recover the momentum vector of each individual atom.After listing the necessary tools to perform this experiment with atoms, the experimental sequence is discussed and the results are presented. In particular we measured a coincidence count reduction that cannot be explained through any simple classical model. This corresponds to the signature of a two-particle interference, and confirms that our atomic pair source produces beams which have highly correlated populations and are well mode matched. This opens the prospect of testing Bell’s inequalities involving mechanical observables of massive particles, such as momentum, using methods inspired by quantum optics. It also demonstrates a new way to produce and benchmark twin-atom pairs that may be of interest for quantum information processing.

Μελετάµε χωρικά εντοπισµένες και χρονικά περιοδικές λύσεις σε διακριτά συστήµατα που εκτείνονται σε µία χωρική διάσταση. Αυτού του είδους οι λύσεις είναι γνωστές µε τον όρο discrete breathers (DB) ή intrinsic localized modes (ILM). Στην ελληνική ϐιϐλιογραϕία, έχουν ονοµαστεί ∆ιακριτές Πνοές. Απαραίτητα χαρακτηριστικά για την εµϕάνιση τέτοιων λύσεων είναι η ύπαρξη ενός άνω φράγµατος του γραµµικού φάσµατος καθώς και η µη γραµµικότητα των εξισώσεων κίνησης, χαρακτηριστικά που συναντάµε σε πολλά φυσικά συστήµατα. Συγκεκριμένα, ασχολούµαστε µε πλέγµατα τύπου Klein Gordon και παρουσιάσουµε μια αποδείξη ύπαρξης τέτοιων λύσεων καθώς και αριθµητικά αποτελέσµατα µελετώντας παράλληλα την ευστάθεια των περιοδικών αυτών λύσεων µέσω της ϑεωρίας Floquet. Πέραν του κλασικού µοντέλου, όπου έχουµε αλληλεπιδράσεις πλησιέστερων γειτόνων, εισάγουµε επίσης ένα νέο µοντέλο µε αλληλεπιδράσεις µακράς εµβέλειας η οποία ελέγχεται µέσω µιας παράµετρου α και µελετάµε τις επιπτώσεις που έχει η μεταβολή του εύρους αλληλεπίδρασης στον χωρικό εντοπισµό και την ευστάθεια ενός DB. / We study time-periodic and spatially localized solutions in discrete dynamical systems describing Hamiltonian lattices in one spatial dimension. These solutions are called discrete breathers (DBs) or intrinsic localized modes (ILM). Necessary conditions for their occurrence are the boundedness of the spectrum of linear oscillations of the system as well as the nonlinearity of the equations of motion. More specifically, we focus on a Klein Gordon lattice and present an existence proof for such solutions, as well as numerical results revealing the stability (or instability) of DBs using Floquet theory. Besides reporting on the classical Klein Gordon model with nearest neighbor interactions, we also introduce long range interactions in our model, which are controlled by a parameter α and study the effect of varying the range of interactions on the spatial localization and the stability of a DB.

Traditionnellement, les applications orientées objets légataires intègrent différents aspects fonctionnels. Ces aspects peuvent être dispersés partout dans le code. Il existe différents types d’aspects : • des aspects qui représentent des fonctionnalités métiers ; • des aspects qui répondent à des exigences non fonctionnelles ou à d’autres considérations de conception comme la robustesse, la distribution, la sécurité, etc. Généralement, le code qui représente ces aspects chevauche plusieurs hiérarchies de classes. Plusieurs chercheurs se sont intéressés à la problématique de la modularisation de ces aspects dans le code : programmation orientée sujets, programmation orientée aspects et programmation orientée vues. Toutes ces méthodes proposent des techniques et des outils pour concevoir des applications orientées objets sous forme de composition de fragments de code qui répondent à différents aspects. La séparation des aspects dans le code a des avantages au niveau de la réutilisation et de la maintenance. Ainsi, il est important d’identifier et de localiser ces aspects dans du code légataire orienté objets. Nous nous intéressons particulièrement aux aspects fonctionnels. En supposant que le code qui répond à un aspect fonctionnel ou fonctionnalité exhibe une certaine cohésion fonctionnelle (dépendances entre les éléments), nous proposons d’identifier de telles fonctionnalités à partir du code. L’idée est d’identifier, en l’absence des paradigmes de la programmation par aspects, les techniques qui permettent l’implémentation des différents aspects fonctionnels dans un code objet. Notre approche consiste à : • identifier les techniques utilisées par les développeurs pour intégrer une fonctionnalité en l’absence des techniques orientées aspects • caractériser l’empreinte de ces techniques sur le code • et développer des outils pour identifier ces empreintes. Ainsi, nous présentons deux approches pour l’identification des fonctionnalités existantes dans du code orienté objets. La première identifie différents patrons de conception qui permettent l’intégration de ces fonctionnalités dans le code. La deuxième utilise l’analyse formelle de concepts pour identifier les fonctionnalités récurrentes dans le code. Nous expérimentons nos deux approches sur des systèmes libres orientés objets pour identifier les différentes fonctionnalités dans le code. Les résultats obtenus montrent l’efficacité de nos approches pour identifier les différentes fonctionnalités dans du code légataire orienté objets et permettent de suggérer des cas de refactorisation. / Object oriented applications integrate various functional aspects. These aspects can be scattered everywhere in the code. There are various types of aspects : • aspects which represent business functionalities ; • aspects related to non functional requirements or to design concerns such as robustness, distribution, and security. The code representing such aspects can be located in different class hierarchies. Researchers have been interested in the problem of the modularisation of these aspects and many approaches were proposed : oriented programming subjects, oriented programming Aspects and oriented programming view. These approaches offer techniques and tools for designing object oriented applications based on the composition of slices of various aspects. The main benefit of the separation of aspects is supporting reuse and maintenance. Consequently, it is well worth identifying and extracting aspects of legacy object oriented applications. Our work mainly focuses on functional aspects. Assuming that the code of a functional aspect or a feature has a functional cohesion (dependencies between elements), we suggest methods for identifying such features from the code. The idea is to identify, in the absence of any aspect oriented paradigm, the techniques used for implementing a feature in the code. Our approach consists of : • identifying techniques used by developers to integrate a feature in the absence of aspect oriented techniques • characterizing the patterns of these techniques • and developing tools to identify these patterns. We present two approaches for the identification of the existing features in the object oriented code. The first one identifies various design patterns which integrates these features in the code. The second approach uses the formal concept analysis to identify the recurring features in the code. We experiment our approaches to identify functional features in different open source object oriented applications. The results show the efficiency of our approaches in identifying various functional features in the legacy object oriented, and can some times suggest refactoring.

La recherche de phases magnétiques exotiques de la matière qui fondent même à T=0 uniquement sous l'action des fluctuations quantiques a été long et ardu, à la fois théoriquement et expérimentalement. La percée est venue récemment avec la découverte de l'Herbertsmithite, un composé formant un réseau kagome parfait avec des moments magnétiques de spin-1/2. Des expériences pionnières, mêlant des mesures de NMR, µSR et de diffusion de neutrons, ont montré une absence totale de gel ou d'ordre des moments magnétiques de spin, fournissant ainsi une forte signature d'une phase paramgnétique quantique. Théoriquement, l'Herbertsmithite est extrêmement bien modélisé par le modèle de Heisenberg quantique antiferromagnétique pour des spins-1/2 sur le réseau kagome, problème qui n'a pas été résolu jusqu'à présent. Plusieurs méthodes approximatives numériques et analytiques ont donné différents états fondamentaux, allant des liquides de spins Z2 gappés et un liquide de spins exotique algébrique U(1) de Dirac aux liquides de spins chiraux et les cristaux à liaisons de valence. Dans cette thèse, le problème est traité dans le cadre d'une approche particule-esclave fermionique, à savoir le formalisme des fermions de Schwinger SU(2). Il est conclu qu'un liquide de spins sans gap algébrique de Dirac a l'énergie variationnelle la plus basse et peut en fait constituer un vrai état fondamental physique de liquide de spins. Une implémentation sophistiquée de méthodes numériques de pointes comme le Monte-Carlo variationnel, le Monte-Carlo fonctions de Green et l'application de pas Lanczos dans un schéma variationnel ont été utilisés. Il est montré que contrairement à la croyance habituelle, le liquide de spins de Dirac U(1) projeté en "2+1" dimensions est remarquablement robuste par rapport à une large classe de perturbations, incluant les liquides de spins topologiques Z2 et les cristaux à liaisons de valence. De plus, l'application de deux pas Lanczos sur la fonction d'onde du liquide de spins de Dirac U(1) montre que son énergie est compétitive avec celles proposées pour les liquides de spins topologiques Z2. Ce résultat, combiné avec les indications expérimentales qui pointent vers un liquide de spins sans gap pour l'Herbertsmithite, appuie l'affirmation que le vrai état fondamental de ce modèle est en fait un liquide de spins algébrique de Dirac.

Η παρούσα εργασία αφορά στη μελέτη Χαμιλτώνιων συστημάτων Ν μη γραμμικών ταλαντωτών, όπως είναι αυτό των Fermi Pasta και Ulam (FPU), με στόχο την βαθύτερη κατανόηση της δυναμικής των σχεδόν-περιοδικών τροχιών και του ρόλου των αντίστοιχων τόρων στο χώρο φάσεων, καθώς αυξάνουμε την ενέργεια Ε και τον αριθμό βαθμών ελευθερίας Ν του συστήματος. Το βασικό μας αποτέλεσμα είναι ότι υπάρχουν τόροι χαμηλής διάστασης, που προκύπτουν από τη συνέχεια των αντίστοιχων του γραμμικού συστήματος, οι οποίοι ευθύνονται για τις FPU επαναλήψεις και εμποδίζουν την ισοκατανομή της ενέργειας μεταξύ όλων των κανονικών τρόπων ταλάντωσης. Αναλύοντας ευστάθεια αυτών των τόρων, μπορέσαμε να δώσουμε μια πληρέστερη ερμηνεία στο Παράδοξο των FPU, συνδέοντας και συμπληρώνοντας έτσι δύο από τις επικρατέστερες ερμηνείες του εν λόγω φαινομένου. / The present work concerns the study of Hamiltonian systems of N nonlinear coupled oscillators, as it is the one by Fermi Pasta and Ulam (FPU), in order to understand the dynamics of quasi-periodic orbits and the role of their corresponding tori in phase space, as we increase the energy E and the number N of the degrees of freedom. Our fundamental result is that there exist tori of low dimension, that come from the continuation of the corresponding tori of the linear system, which are responsible for the FPU recurrences and prevent the system from equipartition of the energy among all normal modes. By investigating the stability of these tori, we achieved to provide a more complete explanation for the FPU paradox, connecting and supplementing in this way two of the most dominant approaches for this paradox.

The advancement in the eld of cold-atoms has generated a lot of interest in the condensed matter community. Cold-atom experiments can simulate clean, disor-der/impurity free systems very easily. In these systems, we have a control over various parameters like tuning the interaction between particles by the Feshbach resonance, tuning the hopping between lattice sites by laser intensity and so on. As a result, these systems can be used to mimic various theoretical models, which was hindered because of various experimental limitations. Thus, we have an ex-perimental tool in which we can start with a simple theoretical model and later tune the model experimentally and theoretically to simulate the real materials. This will be helpful in studying the physics of the real materials as we can control interactions as well as the impurities can also be taken care of. But the advance-ment in the eld of cold atoms has seen a roadblock for the fermions in optical lattices. The super uid and anti-ferromagnetic phases has not been achieved for fermions in optical lattices due to the \cooling problem" (entropy issues). In this thesis, we have addressed the issue of the \cooling problem" for fermions in optical lattice systems and studied the system with determinant quantum Monte Carlo technique. We start by giving a general idea of cold-atoms and optical lat-tice potentials, and a brief review of the experimental work going on in the cold-atomic systems. Experimental limitations like \fermionic cooling problem" have been discussed in some detail. Then we proposed a bilayer band-insulator model to circumvent the \entropy problem" and simultaneously increasing the transi-tion temperature for fermions in optical lattices. We have studied the attractive Hubbard model, which is the minimal model for fermions in optical lattices. The techniques that we have used to study the model are mean- eld theory, Gaussian uctuation theory and determinant quantum Monte Carlo numerical technique. . Chapter-1 : provides a general introduction to the ultra-cold atoms, optical lattice and Feshbach resonance. In this chapter we have discussed about cold-atom experiments in optical lattice systems. Here, we have brie y discussed the control over various parameters in the experiments. The goal of these experiments is to realize or mimic many many-body Hamiltonians in experiments, which until now was just a theoretical tool to describe various many-body physics. In the end we give a brief idea for introducing disorder in the cold-atom experiments discuss the limitations of these experiments in realizing the \interesting" super uid and anti-ferromagnetic phases of fermionic Hubbard model in optical lattices. Chapter-2 : gives a brief idea of \Determinant Quantum Monte-Carlo" (DQM C) technique that has been used to study these systems. In this chapter we will discuss the DQM C algorithm and the observables that can be calculated. We will discuss certain limitation of the DQM C algorithm like numerical instability and sign problem. We will brie y discuss how sign problem doesn't occur in the model we studied. Chapter-3 : discusses the way by which we can bypass the \cooling problem" (high entropy state) to get a fermionic super uid state in the cold atom experi-ments. In this chapter we propose a model whose idea hinges on a low-entropy band-insulator state, which can be tuned to super uid state by tuning the on-site attractive interaction by Feshbach resonance. We show through Gaussian uctua-tion theory that the critical temperature achieved is much higher in our model as compared to the single-band Hubbard model. Through detailed variational Monte Carlo calculations, we have shown that the super uid state is indeed the most stable ground state and there is no other competing order. In the end we give a proposal for its realization in the ultra-cold atom optical lattice systems. Chapter-4 : discusses the DQM C study of the model proposed in chapter- 3. Here we have studied the various single-particle properties like momentum distribution, double occupancies which can be easily measured in cold-atom ex-periments. We also studied the pair-pair and the density-density correlations in detail through DQM C algorithm and mapped out the full T U phase diagram. We show that the proposed model doesn't favor the charge density wave for the interaction strengths we are interested in. Chapter-5 : gives a brief idea of the e ect of adding an on-site random disorder in our proposed bilayer-Hubbard model. We study the e ect of random disorder on various single-particle properties which can be easily veri ed in cold-atom ex-periments. We studied the suppression of the pair-pair correlations as we increase the disorder strength in our proposed model. We nd that the critical value of the interaction doesn't change in the weak-disorder limit. We estimated the critical disorder strength needed to destroy the super uid state and argued that the tran-sition from the super uid to Bose-glass phase in presence of disorder lies in the universality class of (d + 1) XY model. In the end, we give a schematic U V phase diagram for our system. Chapter-6 : We studied the bilayer attractive Hubbard model in different lattice geometry, the bilayer honeycomb lattice, both in presence and in absence of the on-site random disorder. We discussed how the pair-pair and density-density cor-relations behave in the presence and absence of disorder. Through the finite-size scaling analysis we see the co-existence of the super fluid and the charge density wave order at half- lling. An in nitesimal disorder destroys the CDW order com-pletely while the super uid phase found to be robust against weak-disorder. We estimated the critical interaction strength, the critical temperature and the critical disorder strength through nite-size scaling, and provide a putative phase diagram for the system considered.

Abstract An optical lattice is a periodic array produced by the superposition of two or more counter-propagating laser beams which interfere and create a spatial periodic intensity pattern. An important difference between a solid state crystalline lattice and an optically induced lattice is that the latter have a longer coherence time which permits the observation of coherent phenomenon such as Bloch oscillations. Recently, an interesting effect has been reported in the literature, which is the analogue of Rabi oscillations in two level systems driven by an external optical field: an optical beam composed by pair of frequencies, resonants with the high symmetry points of the Brillouin zone, propagating through an optical lattice exhibits oscillations of the Fourier ways between the frequencies due to the coupling with the lattice. Therefore, in the optical case the field plays the role of the matter, while the lattice plays the role of the driving field which couples the frequencies. Based in this result, the objective of this paper is to investigate the influence of self-defocusing in the optical Rabi oscillation. We have shown, through numerical calculation of the wave nonlinear equation, that in the self-defocusing optical lattice, these oscillations go disappearingare. The result of the disappearance of the oscillations of Fourier can be understood considering that self-defocusing promotes an addition of phases in the ways that were resonant - kB and kB, in different way, form that the basic difference between kB and - kB will not be plus a vector of the reciprocal net, and as consequence the oscillations of optic Rabi disappear, therefore the coupling with the net requires this condition in the resonant frequencies. / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Resumo Uma rede opticamente induzida é um arranjo periódico produzido pela superposição de dois ou mais feixes de laser que ao interferirem criam um padrão de intensidade periódico espacial. Uma importante diferença entre uma rede cristalina no estado sólido e uma rede produzida opticamente é que a última apresenta um tempo de coerência longo, permitindo a observação de fenômenos coerentes tal como as oscilações de Bloch. Recentemente, um efeito interessante análogo às oscilações de Rabi em um sistema atômico de dois níveis impulsionado por um campo externo, foi publicado na literatura: um feixe óptico composto de um par de freqüências, ressonantes com os pontos de alta simetria da zona de Brillouin, propagando - se através de uma rede óptica exibe oscilações dos modos de Fourier entre as freqüências adequadas para o acoplamento com a rede. Portanto, no caso óptico o campo representa a matéria, enquanto a rede representa a intensidade do campo aplicado que acopla as freqüências. Baseado neste resultado, o objetivo desta dissertação é investigar a influência da auto - desfocalização nas oscilações ópticas de Rabi. Demonstraremos, através de cálculos numéricos da equação de onda não linear, que em uma rede opticamente induzida na presença da auto - desfocalização, essas oscilações vão desaparecendo. O resultado do desaparecimento das oscilações de Fourier pode ser compreendidoconsiderando que a auto - desfocalização promove um adicionamento de fases nos modos que eram ressonantes -kB e kB, de maneira diferente, de forma que a diferença básica entre kB e -kB não será mais um vetor da rede recíproca, e como conseqüência as oscilações de Rabi óptica desaparecem, pois o acoplamento com a rede requer esta condição nas freqüências ressonantes.