Neste trabalho estudamos uma rede hexagonal com uma cadeia de impurezas nas bordas e com supercondutividade induzida, de forma a mostrar a existência de fases com férmions de Majorana. Para tal, começamos introduzindo invariantes topológicos, número de Chern e Z2 e mostramos dois modelos para rede hexagonal. O primeiro, modelo de Haldane, fazemos como motivação histórica. O segundo, modelo de Kane-Mele, é usado como base para todo o trabalho. Seguimos introduzindo supercondutividade e como ela ocorre quando aplicada junto do Modelo de Kane-Mele, o método auto-consistente e quais as condições necessárias para termos supercondutividade apenas nas bordas. Continuamos com efeitos de impurezas magnéticas nas bordas e introduzimos férmions de Majorana que são os alvos principais dos resultados. Mostramos então, que existe fases topológicas em cadeias de impureza magnética, com momentos em espiral, contudo o diagrama de fase depende de várias condições. Por fim, mostramos que a variação da fase topológica se deve a oscilações nos níveis de energia em que o invariante topológico também varia, contrariando resultados obtidos para a rede quadrada. Concluímos esse trabalho com implicações experimentais desse resultado e possíveis caminhos que podem ser seguidos. / In this work, we study a honeycomb lattice with induced superconductivity and edge impurity in order to show the existence of a phase that host Majorana bound state. To do so, we start introducing topological invariants, Chern number and Z2, and we show two models for honeycomb lattice. The first, Haldane\'s Model, due its historical importance. The second, Kane-Mele model, because it will be used during all this work. Then we review superconductivity, showing the self-consistent method, and we apply it to Kane-Mele model, in which we find some necessary conditions to induce superconductivity only at the edges. From this point, we study the effect of magnetic impurities at the edges, and we introduce Majorana bound state, that will be the main objective of our results. In our results, we show the existence of topological non-trivial phases for spiral magnetic chain in the zigzag edge. With this we make a phase diagram. We also find oscillation in the energy spectrum and the topological phase changes with the oscillation, this is different from square lattice in which we should not have a change in the topological phase. We conclude this work with experimental implications of our result and possible developments.

O objetivo deste trabalho visa a prática da psicanálise lacaniana e sua fundamentação, tendo a civilização como parceira. Os impasses da civilização do risco e suas incidências sobre o corpo interessam-nos como viés para uma reflexão sobre a prática da psicanálise lacaniana na atualidade, especialmente em sua relação com os novos sintomas, sobretudo a partir do início do séc. XXI. O interesse em dialogar com a contemporaneidade visa fundamentar ainda mais a experiência da psicanálise aplicada, razão da forte presença dos psicanalistas nas instituições. Não nos detivemos apenas em aspectos da contemporaneidade. Para melhor situar na prática lacaniana a noção de falasser [parlêtre], a partir do último ensino de Lacan, resgatamos o debate sobre os universais, a aposta de Pascal e o mito hegeliano do senhor/mestre e do escravo. / This paper is related to the Lacanian psychoanalytical practice and its theoretical fundaments based on civilization as support . Civilization impasses on risk and their incidences on the body are considered as they lead to a reflection about the practice of Lacanian psychoanalysis in our days, especially when new symptoms are concerned, since the beginning of the 21st century. The interest in sustaining, with our contemporary times, a dialogue aims to add fundaments to applied psychoanalysis, considering the relevant presence of psychoanalysts in the institutions. This paper is nor more limited to what is found in contemporary times to better situate, in the Lacanian practice, the concept of parlêtre (by letter made) from Lacans last teaching. We have also recovered the debate about universals, Pascals bet and the Hegelian myth about the master and the slave.

Submitted by João Valente (joaopaulovalente@hotmail.com) on 2013-09-21T14:35:40Z No. of bitstreams: 1 Tese_FINAL.pdf: 645576 bytes, checksum: 5dcd3933e691c9eb4d85777d7f20904f (MD5) / Approved for entry into archive by Janete de Oliveira Feitosa (janete.feitosa@fgv.br) on 2013-09-27T13:44:20Z (GMT) No. of bitstreams: 1 Tese_FINAL.pdf: 645576 bytes, checksum: 5dcd3933e691c9eb4d85777d7f20904f (MD5) / Approved for entry into archive by Marcia Bacha (marcia.bacha@fgv.br) on 2013-09-30T12:35:01Z (GMT) No. of bitstreams: 1 Tese_FINAL.pdf: 645576 bytes, checksum: 5dcd3933e691c9eb4d85777d7f20904f (MD5) / Made available in DSpace on 2013-09-30T12:35:14Z (GMT). No. of bitstreams: 1 Tese_FINAL.pdf: 645576 bytes, checksum: 5dcd3933e691c9eb4d85777d7f20904f (MD5) Previous issue date: 2013-06-19 / Neste trabalho, propusemos um modelo DSGE que busca responder algumas questões sobre políticas de afrouxamento monetário (Quantitative Easing - QE) recentemente implementadas em resposta à crise de 2008. Desenvolvemos um modelo DSGE com agentes heterogêneos e preferred-habitat nas compras de títulos do governo. Nosso modelo permite o estudo da otimalidade da compra de portfolio (em termos de duration dos títulos) para os bancos centrais quando estão implementando a política. Além disso, a estrutura heterogênea nos permite olhar para distribuição de renda provocada pelas compras de títulos. Nossos resultados preliminares evidenciam o efeito distributivo do QE. No entanto, nosso modelo expandido apresentou alguns problemas de estabilidade. / In this paper, we proposed a DSGE model that seeks to answer some questions about the recent implemented Quantitative Easing (QE) programs. Our framework is a DSGE model with heterogeneous agents and preferred-habitat in purchases of government bonds. It allows the study of optimality purchasing portfolio (in terms of duration of the bonds) for central banks when they are implementing the policy. Furthermore, the heterogeneous structure allows us to look at income distribution caused by purchases of these securities. Our preliminary results show some distributive effect of QE. However, our expanded model showed some stability problems.

Le pronom-adverbe y en espagnol ancien a disparu à la fin du XVe siècle. Il a survécu néanmoins à travers les signes hay, soy, estoy, doy et voy. L’objet de la présente thèse est d’expliquer les raisons de sa disparition, d’une part, et de sa survivance dans ces cinq formes verbales, d’autre part. Elle se divise en deux parties. La première porte sur l’analyse du morphème stématique y et implique l’étude du système de représentation de l’espace. La seconde concerne l’étude des formes verbales dans lesquelles on retrouve ce morphème sous forme astématique et implique une approche sémantique des verbes existentiels. En fonction de la façon dont ses usagers perçoivent le monde, chaque langue élabore un système de représentation de l’espace, qui lui est propre. Dans la langue espagnole deux systèmes s’offraient aux locuteurs médiévaux : l’un représenté par les anaphoriques y et ende, l’autre par les déictiques aquí, acá, ahí, allí, allá, acullá, aquende et allende. Ces deux systèmes – qui se sont opposés jusqu’à la disparition de l’un d’eux, –représentaient une conceptualisation de l’espace différente. Dans le premier de ces systèmes y signifie l’espace indivis. Le deuxième, en revanche, divise l’espace en fonction du MOI locuteur et du lieu que celui-ci occupe. La sémiologie des formes verbales, quant à elle, montre un ensemble spécifique dont les éléments sont marqués dans leur signifiant par ce -y et dont la cohésion est assurée par la sémantèse des verbes existentiels. L’hypothèse présentée ici va à l’encontre de la théorie de la subduction, défendue en particulier par Maurice Molho, qui fonde la coalescence des quatre premières formes verbales sur analogie avec hay. / The adverbial pronoun y vanished from Old Spanish at the end of the XVth century. It survived, however, in the signs hay, soy, estoy, doy et voy. The purpose of the present doctoral dissertation is to explain why the pronoun as such disappeared , in the first place, and why, on the other hand, it did survive in the five above-mentioned verbal forms. The present work is comprised of two sections. The first one analyses the stematic morpheme y, a task which implies studying the system of spatial representations. The second one deals with the verbal forms containing this same morpheme in astematic form and this in turn implies a semantic approach of existential verbs. Every language generates its own system of spatial representation, depending on how its users perceive the world. In Spanish, two different systems were available to the medieval speakers: one represented by the anaphoric elements y and ende, the other one represented by the deictic words aquí, acá, ahí, allí, allá, acullá, aquende and allende. These two systems were harbouring two different conceptualizations of space. In the first system, y signifies space as a whole. The second one, on the contrary, splits the space in relation to the EGO as SPEAKER and to the locus of that EGO. The semiology of the five verbal forms delineates a specific set of elements, the signifiers of which are marked by -y and whose general consistency is based on the semantesis of existential verbs. The hypothesis presented in the present work runs contrary to the subduction theory, advocated in particular by Maurice Molho, which explains the coalescence in the first four verbal forms through an analogy with hay.

Context: Most of the problem of generating timetable for a school belongs to the class of NP-hard problems. Complexity and practical value makes this kind of problem interesting for parallel computing. Objectives: This paper focuses on Class-Teacher problem with weighted time slots and proofs that it is NP-complete problem. Branch and bound scheme and two approaches to distribute the simulated annealing are proposed. Empirical evaluation of described methods is conducted in an elementary school computer laboratory. Methods: Simulated annealing implementation described in literature are adapted for the problem, and prepared for execution in distributed systems. Empirical evaluation is conducted with the real data from Polish elementary school. Results: Proposed branch and bound scheme scales nearly logarithmically with the number of nodes in computing cluster. Proposed parallel simulated annealing models tends to increase solution quality. Conclusions: Despite a significant increase in computing power, computer laboratories are still unprepared for heavy computation. Proposed branch and bound method is infeasible with the real instances. Parallel Moves approach tends to provide better solution at the beginning of execution, but the Multiple Independent Runs approach outruns it after some time. / Sammanhang: De flesta problem med att generera scheman för en skola tillhör klassen av NP-svårt problemen. Komplexitet och praktiskt värde gör att den här typen av problemen forskas med särskild uppmärksamhet på en parallell bearbetning.   Syfte: Detta dokument fokusarar på Klass-Lärare problem med vikter för enskilda tidsluckor och på att visa var ett NP-svårt problem är fullständigt. Branch and bound scheman och två metoder för att distribuera en simulerad glödgning algoritm presenterades. En empirisk analys av beskrivna metoder gjordes i datorlaboratorium i en grundskola. Metod: Implementering av en simulerad glödgning algoritm som beskrivs i litteraturen blev anpassad till ett utvalt problem och distribuerade system. Empirisk utvärdering genomförs med verkliga data från polska grundskolan Resultat: Föreslagit Branch and bound system graderar nästan logaritmiskt antal noder i ett datorkluster. Den simulerade glödgning algoritmen som föreslagits förbättrar lösningarnas kvalitet. Slutsatser: Trots att en betydande ökning med beräkningskraft är inte datasalar i skolor anpassad till avancerade beräkningar. Användning av den Branch and Bound föreslagna metoden till praktiska problem är omöjlig i praktiken. En annan föreslagen metod Parallel Moves ger bättre resultat i början av utförandet men Multiple Independent Runs hittar bättre lösningar efter en viss tid.

Cette thèse concerne des propriétés des variétés jacobiennes de courbes de genre 2 qui couvrent des courbes elliptiques. Soit E une courbe plane, donnée par une équation y^2=F(x), où F(x)=x^3+a2x^2+a1x+a0 est un polynôme à coefficients rationnels, qui a trois racines distinctes. Pour des raisons historiques, une telle courbe est appelée courbe elliptique. On sait que toute courbe elliptique E peut être équipée d'une structure de groupe commutatif - on peut additionner et soustraire ses points. Un point O « à l'infini », qui est contenu dans toutes les droites verticales (droites de la forme x=c), est l'élément neutre. Cette structure de groupe est décrite par la condition que trois points P,Q,R sur E satisfont P+Q+R=O si et seulement s'ils sont alignés. Les surfaces avec une structure de groupe commutatif sont appelées abéliennes. Par exemple, un produit de deux courbes elliptiques E1xE2 est une surface abélienne, de façon évidente. Considérons maintenant une courbe plane C donnée par une équation y^2=G(x), où G(x)=x^6+b5x^5+b4x^4+b3x^3+b2x^2+b1x+b0 est un polynôme à coefficients rationnels, qui a six racines distinctes. La courbe C est appelée hyperelliptique et n'a pas de structure de groupe. Par contre, nous pouvons lui associer, d'une façon naturelle, une surface abélienne Jac(C), appelée la jacobienne de C. En plus, nous pouvons plonger C dans Jac(C). Certaines courbes hyperelliptiques sont spéciales car elles couvrent des courbes elliptiques. Par exemple, considérons une courbe C donnée par l'équation y^2=x^6+ax^4+bx^2+c, dans laquelle seulement des puissances paires de x apparaissent. Si (x,y) est un point de cette courbe alors de même (-x,y), et nous pouvons définir une application algébrique f:(x,y)->(x^2,y) de degré 2, c'est-à-dire, de fibre générale à deux points. Alors (X,Y)=(x^2,y) est un point de la courbe elliptique E donnée par Y^2=X^3+aX^2+bX+c et nous disons que C est un revêtement double de E. Si E1 est une courbe elliptique, si C est une courbe hyperelliptique, et si C->E1 est un revêtement de degré n qui n'est pas une composition de revêtements, alors nous pouvons plonger E1 dans la surface Jac(C) comme un sous-groupe. De plus, il existe une autre courbe elliptique E2 et un revêtement C->E2 de degré n, tel que la surface Jac(C) a une propriété spéciale - elle peut être obtenue comme quotient de la surface E1xE2 par un sous-groupe fini. Le chapitre 1 de cette thèse traite les aspects géométriques de cette situation. Nous cherchons à savoir quelles courbes peuvent avoir une telle relation et nous nous concentrons surtout sur les cas n=2 et n=3, qui ont déjà été analysés dans la littérature. Dans le cas général, nous obtenons quelques résultats, mais une description complète s'avère très difficile de manière explicite. Le chapitre 2 traite les aspects arithmétiques de la situation, via la théorie des fonctions hauteurs, qui sont un outil très utile pour répondre à des questions concernant des points rationnels de courbes et surfaces. Pour tout nombre rationnel x=a/b, avec a et b des entiers premiers entre eux, on définit la hauteur h(x) de x, de façon très précise, comme une mesure de sa complexité arithmétique - la hauteur dit approximativement combien de chiffres sont nécessaires pour écrire les entiers a et b. De la même façon, la hauteur d'un point rationnel d'une courbe ou surface nous dit combien de chiffres ont les coordonnées. Par exemple, (3,5) et (1749/1331,-1861/1331) sont deux points rationnels de complexités plutôt différentes de la courbe y^2=x^3-x+1, tandis que (2,√7) n'est pas un point rationnel. Il est possible d'attacher une hauteur aux courbes elliptiques et aux surfaces abéliennes qui mesure leur complexité arithmétique totale. Une relation spécifique entre ces deux notions de hauteur est alors conjecturée et nous étudions cette conjecture dans la situation décrite plus haut. Nous montrons que cette relation est vraie pour E1xE2 si et seulement si elle est vraie pour Jac(C). / This thesis deals with properties of Jacobians of genus two curves that cover elliptic curves. Let E be a curve in the plane, given by an equation y^2=F(x), where F(x)=x^3+a2x^2+a1x+a0 is a polynomial with rational coefficients and with three distinct roots. For historical reasons, such a curve is known as an elliptic curve. It is known that every elliptic curve E can be equipped with a structure of a commutative group - its points can be added and subtracted. A point O "at infinity", which is contained in all vertical lines (lines of form x=c), is the neutral element. This group structure is described by the condition that three points P,Q,R in E satisfy P+Q+R=O if and only if they are collinear. Surfaces with a commutative group structure are called abelian. For example, a product of two elliptic curves E1xE2 is an abelian surface in the obvious way. Next we consider a planar curve C given by an equation y^2=G(x), where G(x)=x^6+b5x^5+b4x^4+b3x^3+b2x^2+b1x+b0 is a polynomial with rational coefficients and six distinct roots. The curve C is called hyperelliptic and it does not have a group structure. However, we can associate to it, in a natural way, an abelian surface Jac(C), called the Jacobian of C. Moreover, we can embed C into it. Some hyperelliptic curves, of the form y^2=G(x) as above, are special because they cover elliptic curves. For example, consider a curve C given by y^2=x^6+ax^4+bx^2+c, so that only even powers of x appear. If (x,y) is a point on this curve then so is (-x,y) and we can define an algebraic map f:(x,y)->(x^2,y), that is of degree 2, i.e. 2-to-1. Now (X,Y)=(x^2,y) is a point on the elliptic curve E given by Y^2=X^3+aX^2+bX+c and we say that C is a double cover of E. If E1 is an elliptic curve, C is a hyperelliptic curve, and C->E1 is an n-to-1 covering that is not a composition of coverings, then we can embed E1 into the surface Jac(C) as a subgroup. Moreover, there exists another elliptic curveE2 and an n-to-1 covering C->E2, such that the surface Jac(C) has a special property - it can be obtained as the quotient of the surface E1xE2 by a finite subgroup. The first chapter of the thesis deals with the geometric aspects of this setup. We investigate which curves can form this special relationship and we focus mostly on the cases n=2 and n=3, which have already been analysed in literature. We also gain some insight into the general case, but a full description proves to be very difficult computationally. The second chapter deals with the arithmetic aspects of the setup, via the theory of height functions, which are a very useful tool in answering questions about rational points on curves and surfaces. For every rational number x=a/b, where a and b are coprime integers, one can define its height h(x), in a very precise way, as a measurement of its arithmetic complexity - the height roughly tells us how many digits are needed to write down the integers a and b. Likewise, the height of a rational point on a curve or surface tells us about the number of digits of the coordinates. For example, (3,5) and (1749/1331,-1861/1331) are two rational points of rather different complexity on the curve y^2=x^3-x+1, while (2,√7) is not a rational point. It is also possible to associate a height to an elliptic curve or an abelian surface and measure its arithmetic complexity as a whole. A specific relation between these two heights is conjectured and we investigate it in the context of the setup above. We show that this relation holds for E1xE2 if and only if it holds for Jac(C).

Abstract High computational complexity constrains truly exhaustive computer searches for good space-time (ST) coded modulations mostly to low constraint length space-time trellis codes (STTrCs). Such codes are primarily devised to achieve maximum transmit diversity gain. Due to their low memory order, optimization based on the design criterion of secondary importance typically results in rather modest coding gains. As another disadvantage of limited freedom, the different low memory order STTrCs are almost exclusively constructed for either slow or fast fading channels. Therefore in practical applications characterized by extremely variable Doppler frequencies, the codes typically fail to demonstrate desired robustness. On the other hand, the main drawback of eventually increased constraint lengths is the prohibitively large decoding complexity, which may increase exponentially if optimal maximum-likelihood decoding (MLD) is applied at the receiver. Therefore, robust ST coded modulation schemes with large equivalent memory orders structured as to allow sub-optimal, low complexity, iterative decoding are needed. To address the aforementioned issues, this thesis proposes parallel concatenated space-time turbo coded modulation (STTuCM). It is among the earliest multiple-input multiple-output (MIMO) coded modulation designs built on the intersection of ST coding and turbo coding. The systematic procedure for building an equivalent recursive STTrC (Rec-STTrC) based on the trellis diagram of an arbitrary non-recursive STTrC is first introduced. The parallel concatenation of punctured constituent Rec-STTrCs designed upon the non-recursive Tarokh et al. STTrCs (Tarokh-STTrCs) is evaluated under different narrow-band frequency flat block fading channels. Combined with novel transceiver designs, the applications for future wide-band code division multiple access (WCDMA) and orthogonal frequency division multiplexing (OFDM) based broadband radio communication systems are considered. The distance spectrum (DS) interpretation of the STTuCM and union bound (UB) performance analysis over slow and fast fading channels reveal the importance of multiplicities in the ST coding design. The modified design criteria for space-time codes (STCs) are introduced that capture the joint effects of error coefficients and multiplicities in the two dimensional DS of a code. Applied to STTuCM, such DS optimization resulted in a new set of constituent codes (CCs) for improved and robust performance over both slow and fast fading channels. A recursive systematic form with a primitive equivalent feedback polynomial is assumed for CCs to assure good convergence in iterative decoding. To justify such assumptions, the iterative decoding convergence analysis based on the Gaussian approximation of the extrinsic information is performed. The DS interpretation, introduced with respect to an arbitrary defined effective Hamming distance (EHD) and effective product distance (EPD), is applicable to the general class of geometrically non-uniform (GNU) CCs. With no constrains on the implemented information interleaving, the STTuCM constructed from newly designed CCs achieves full spatial diversity over quasi-static fading channels, the condition commonly identified as the most restrictive for robust performance over a variety of Doppler spreads. Finally, the impact of bit-wise and symbol-wise information interleaving on the performance of STTuCM is studied.

Cette thèse se concentre sur l'étude d'un algorithme multi niveaux de régions de confiance en norme infinie, conçu pour la résolution de problèmes d'optimisation non linéaires de grande taille pouvant être soumis a des contraintes de bornes. L'étude est réalisée tant sur le plan théorique que numérique. L'algorithme RMTR8 que nous étudions ici a été élaboré a partir de l'algorithme présente par Gratton, Sartenaer et Toint (2008b), et modifie d'abord en remplaçant l'usage de la norme Euclidienne par une norme infinie, et ensuite en l'adaptant a la résolution de problèmes de minimisation soumis a des contraintes de bornes. Dans un premier temps, les spécificités du nouvel algorithme sont exposées et discutées. De plus, l'algorithme est démontré globalement convergent au sens de Conn, Gould et Toint (2000), c'est-a-dire convergent vers un minimum local au départ de tout point admissible. D'autre part, il est démontre que la propriété d'identification des contraintes actives des méthodes de régions de confiance basées sur l'utilisation d'un point de Cauchy peut être étendue a tout solveur interne respectant une décroissance suffisante. En conséquence, cette propriété d'identification est aussi respectée par une variante particulière du nouvel algorithme. Par la suite, nous étudions différents critères d'arrêt pour les algorithmes d'optimisation avec contraintes de bornes afin de déterminer le sens et les avantages de chacun, et ce pour pouvoir choisir aisément celui qui convient le mieux a certaines situations. En particulier, les critères d'arrêts sont analyses en termes d'erreur inverse (backward erreur), tant au sens classique du terme (avec l'usage d'une norme produit) que du point de vue de l'optimisation multicritères. Enfin, un algorithme pratique est mis en place, utilisant en particulier une technique similaire au lissage de Gauss-Seidel comme solveur interne. Des expérimentations numériques sont réalisées sur une version FORTRAN 95 de l'algorithme. Elles permettent d'une part de définir un panel de paramètres efficaces par défaut et, d'autre part, de comparer le nouvel algorithme a d'autres algorithmes classiques d'optimisation, comme la technique de raffinement de maillage ou la méthode du gradient conjugue, sur des problèmes avec et sans contraintes de bornes. Ces comparaisons numériques semblent donner l'avantage à l'algorithme multi niveaux, en particulier sur les cas peu non-linéaires, comportement attendu de la part d'un algorithme inspire des techniques multi grilles. En conclusion, l'algorithme de région de confiance multi niveaux présente dans cette thèse est une amélioration du précédent algorithme de cette classe d'une part par l'usage de la norme infinie et d'autre part grâce a son traitement de possibles contraintes de bornes. Il est analyse tant sur le plan de la convergence que de son comportement vis-à-vis des bornes, ou encore de la définition de son critère d'arrêt. Il montre en outre un comportement numérique prometteur. / This thesis concerns the study of a multilevel trust-region algorithm in infinity norm, designed for the solution of nonlinear optimization problems of high size, possibly submitted to bound constraints. The study looks at both theoretical and numerical sides. The multilevel algorithm RMTR8 that we study has been developed on the basis of the algorithm created by Gratton, Sartenaer and Toint (2008b), which was modified first by replacing the use of the Euclidean norm by the infinity norm and also by adapting it to solve bound-constrained problems. In a first part, the main features of the new algorithm are exposed and discussed. The algorithm is then proved globally convergent in the sense of Conn, Gould and Toint (2000), which means that it converges to a local minimum when starting from any feasible point. Moreover, it is shown that the active constraints identification property of the trust-region methods based on the use of a Cauchy step can be extended to any internal solver that satisfies a sufficient decrease property. As a consequence, this identification property also holds for a specific variant of our new algorithm. Later, we study several stopping criteria for nonlinear bound-constrained algorithms, in order to determine their meaning and their advantages from specific points of view, and such that we can choose easily the one that suits best specific situations. In particular, the stopping criteria are examined in terms of backward error analysis, which has to be understood both in the usual meaning (using a product norm) and in a multicriteria optimization framework. In the end, a practical algorithm is set on, that uses a Gauss-Seidel-like smoothing technique as an internal solver. Numerical tests are run on a FORTRAN 95 version of the algorithm in order to define a set of efficient default parameters for our method, as well as to compare the algorithm with other classical algorithms like the mesh refinement technique and the conjugate gradient method, on both unconstrained and bound-constrained problems. These comparisons seem to give the advantage to the designed multilevel algorithm, particularly on nearly quadratic problems, which is the behavior expected from an algorithm inspired by multigrid techniques. In conclusion, the multilevel trust-region algorithm presented in this thesis is an improvement of the previous algorithm of this kind because of the use of the infinity norm as well as because of its handling of bound constraints. Its convergence, its behavior concerning the bounds and the definition of its stopping criteria are studied. Moreover, it shows a promising numerical behavior.

[pt] A Programação Quadrática Binária Irrestrita (UBQP) é amplamente estudada. Trata-se de uma ferramenta de modelagem poderosa, mas otimizar de um problema NP-difícil. Neste trabalho uma nova abordagem é apresentada, que pode ser usada para construir um algoritmo exato. Além disso, a ideia básica que fundamenta o trabalho pode ser usado em um espectro ainda mais amplo de problemas. O algoritmo exato derivado do novo método é altamente paralelizável, o que é uma característica desejável nos dias de hoje em que cloud computing já é uma realidade. Para instâncias razoavelmente grandes do UBQP, o novo método pode paralelizar a centenas, ou até milhares, de núcleos com facilidade, com um aumento de desempenho quase linear. / [en] Unconstrained Binary Quadratic Programming (UBQP) is widely studied. It is a powerful modeling tool and its associate problem is NP-hard. In this work a new approach is introduced, which can be used to build an exact algorithm. Also, the fundamental idea behind it can be used in an even wider family of problems. This exact algorithm derived from the new method is highly parallelizable, which is a desired feature nowadays, when the cloud computing is a reality. For reasonably large instances of UBQP, the new method can parallelize to hundreds, or even thousands, of cores easily, with a near-linear speedup.

Le thème central de cette thèse porte sur des partitions en n parties de l'intervalle entier [1, N] = {1,2,...,N} excluant la présence, dans chaque partie, de solution de l'équation x + y = z dans le cas classique, ou seulement de telles solutions avec x ≠ y dans le cas faible. Pour n donné, le plus grand N admissible dans le cas classique se note S(n) et s'appelle le n-ème nombre de Schur ; dans le cas faible, il se note WS(n) et s'appelle le n-ème nombre de Schur faible. Bien qu'introduits il y a plusieurs décénnies déjà, et même il y a un siècle dans le cas classique, on ne sait encore que très peu de choses au sujet de ces nombres. En particulier, S(n) et WS(n) ne sont exactement connus que pour n ≤ 4. Cette thèse est composée de deux chapitres : le premier revisite des encadrements connus sur les nombres de Schur classiques et faibles, et le second est consacré à la construction de nouveaux minorants des nombres de Schur faibles WS(n) pour n = 7, 8 et 9. Nous introduisons, dans le premier chapitre, les ensembles t-libres de sommes, t ∊ ℕ, dont l'utilisation permet de généraliser et d'unifier diverses démonstrations de majorants des S(n) et WS(n). Nous obtenons également une relation entre WS(n + 1) et WS(n). Dans le second chapitre, nous initions l'étude de certaines partitions hautement structurées présentant un potentiel intéressant pour le problème de minorer les nombres WS(n). Effectivement, avec des algorithmes de recherche ne portant que sur ces partitions, nous retrouvons les meilleurs minorants connus sur WS(n) pour 1 ≤ n ≤ 6, et nous améliorons significativement ceux pour 7 ≤ n ≤ 9. / The main theme of this thesis is about partitions in n parts of the integer interval [1, N] = {1,2,...N} excluding the presence, in each part, of solutions of the equation x + y = z in the classical case, or only of such solution with x ≠ y in the weak case. For given n, the largest admissible N in the classical case, it is denoted S(n) and called the n-th Schur number ; in the weak case, it is denoted WS(n) and called the n-th weak Schur number. Even though these numbers were already introduced several decades ago, and even a century ago in the classical case, almost nothing is known about them. In particular, S(n) and WS(n) are exactly known for n ≤ 4. This thesis comprises two chapters :the first one revisits known lower and upper bounds on the classical weak Schur numbers, and the second one is dedicated to the construction of the new lower bounds on the weak Schur numbers WS(n) for n = 7,8 and 9. In the first chapter, we introduce the t-sumfree sets, t ∊ ℕ, which allow us to generalize and unify various proofs concerning upper bounds on S(n) and WS(n). We also obtain a new relationship between WS(n + 1) and WS(n).In the second chapter, we initiate the study of certain highly structured partitions which present an interesting potential for the problem of bounding the numbers WS(n) from below. Indeed, with search algorithms considering only partitions, we rediscover the best known lower bounds on WS(n) for 1 ≤ n ≤ 6, and we significatively improve those for 7 ≤ n ≤ 9.