This technical report complements the paper entitled “Scalable error isolation for distributed systems” published at USENIX NSDI 15.

In the present work, we study some aspects of Einstein's theory of gravitation in general spacetimes with an arbitrary number of dimensions. In the first chapter we summarize the foundations of used geometric formalism and we derive the equation of goedesic deviation representing the relation between relative acceleration and the Riemann tensor. Second chapter presents different types of algebraic classification of the Weyl tensor in four and higher dimensions. Third chapter is devoted to a detailed examination of the test particle motions and also to the interpretation of different terms in the general equation of geodesic deviation. The fourth section examines appropriate choice of the interpretation frame and the coordinates. The final fifth chapter contains an analysis of the motion of test particles in the Robinson-Trautman spacetime with an arbitrary higher number of dimensions.

Cette thèse fournit une analyse de la probabilité, avec une considération particulière du rôle que jouent les symétries et l’invariance dans son caractère objectif. La thèse défend un dualisme rationnel-physique. Nous développons une théorie de la probabilité épistémique ainsi qu’une théorie de la probabilité physique. La première concerne les degrés de croyance rationnels ; la seconde, la propension singulière peu fluctuante sur laquelle émergent les fréquences relatives stables. Du côté épistémique, nous défendons le bayésianisme objectif et ses règles d’attribution de probabilité, l’attribution invariante et la maximisation d’entropie. Nous généralisons également le bayésianisme orthodoxe et sa règle de changement de probabilité, la conditionnalisation, à la minimisation de la divergence Kullback-Leibler. Le bayésianisme orthodoxe généralisé est développé à partir d’une analyse générale de l’apprentissage, incluant la théorie AGM et la théorie de rang. L’analyse de l’opposition des deux bayésianismes culmine dans un pluralisme du bayésianisme combiné, instancié par une famille de révisions probabilistes qui répondent au problème de l’itération. Du côté physique, nous développons une explication de la fréquence relative à partir de l’approche par la loi des grands nombres. Nous répondons au dilemme de Gillies, selon lequel une théorie scientifique objective de la propension singulière et de long terme est impossible. Dans ce cadre, nous développons la méthode des fonctions arbitraires comme attribution de propension singulière peu fluctuante, et proposons une analyse détaillée de la mécanique statistique et du cas paradigmatique du lancer de pièce. / This thesis analyses the concept of probability and the role of symmetry and invariance in its objectif character. It defends a rational-physical dualism. I first develop a theory of epistemic probability, which addresses (the rational degrees of belief. I also develop a theory of physical probability, conceived as single case propensity on which stable frequencies emerge. Epistemically, I defend an objective Bayesianism and its rules of probability attribution, that is, the invariant prior attribution and maximizing entropy. I also generalize orthodox Bayesianism and its rule of probability change, conditionalization, to the minimization of the Kullback-Leibler divergence. Generalized orthodox Bayesianim is developed from a general investigation of learning, which includes the AGM theory and ranking theory. I resolve the opposition of the two Bayesianims through a pluralism of combined Bayesianism, instantiated by a family of probabilistic revisions which solve the iteration problem. Physically, I explain stable relative frequencies with the law of large numbers approach. I answer Gillies dilemma, according to which there is no scientific objective theory of propensity that is both single case and long term. I here develop the method of arbitrary functions as an attribution of relatively stable single case propensity and analyse in detail statistical mechanics and the paradigmatic coin toss.

Les ions de terres rares en matrice cristalline, refroidis à très basse température, offrent des propriétés remarquables pour le traitement analogique du signal sur porteuse optique. L’élargissement inhomogène du spectre d’absorption peut en effet atteindre plusieurs centaines de gigahertz alors que la largeur homogène des raies d’absorption des ions individuels ne dépasse pas quelques kilohertz. Par pompage optique il est alors possible de modifier à volonté le profil du spectre d’absorption. On dispose ainsi d’un filtre programmable présentant à la fois une très grande bande passante, donnée par la largeur inhomogène, et une excellente résolution, fixée par la largeur homogène. Une raie d’absorption étroite correspond à un état de superposition quantique de longue durée de vie. C’est sous cet angle, celui des transitoires cohérents, et spécifiquement celui des échos de photons que nous abordons les propriétés du filtre programmable. Dans la première partie de la thèse, le filtre est programmé comme un élément dispersif. Il permet d’atteindre des taux de dispersion inaccessibles aux dispositifs optiques conventionnels, tels que les fibres optiques. Nous l’utilisons comme un composant de lentille temporelle, en vue de produire des signaux de forme arbitraire. Par rapport à des dispositifs d’optique conventionnels, on gagne plusieurs ordres de grandeurs en termes de produit temps x bande passante. Après avoir exploité l’écho de photon dans un contexte de filtrage linéaire, nous tirons parti de ses propriétés de très forte non-linéarité dans la seconde partie de la thèse. Cette fois nous cherchons à capturer un signal lumineux de très faible intensité, à le convertir en état de superposition atomique, puis à le restituer dans son état lumineux initial. Cela suppose en particulier d’empêcher les effets d’émission spontanée ou stimulé qui nuisent à la fidélité de la restitution. Pour ce faire, nous proposons un nouveau protocole que nous avons appelé « Revival Of Silenced Echo » (ROSE) / Rare earth ions doped crystals, when cooled at very low temperature, exhibit outstanding properties for optically-carried analogical signal processing. The absorption spectral broadening can reach several hundred of Gigahertz, while the homogeneous width of each individual ion does no exceed a few kilohertz. With the help of optical pumping, one may modify the absorption profile at will. The resulting programmable filter simultaneously offers a very large bandwidth, given by the inhomogeneous width, and a very good resolution, fixed by the homogeneous width. Narrow absorption line is related to long lifetime quantum superposition. We contemplate the programmable filter properties, keeping in mind this coherent transient picture, specifically related to photon echoes. In the first part of the dissertation, the programmable filter is programmed as a dispersive element. This gives access to dispersion rate values out of reach of conventional optical devices, such as optical fibers. We use the filter as a temporal lens component, with an eye to generating arbitrary waveforms. Thereby, we gain several orders of magnitude against conventional optical devices in terms of time x bandwidth product. After taking advantage of photon echoes in the linear filtering context, we capitalize on their strongly non-linear properties in the second part of the dissertation. This time we want to capture a very weak optical signal, to convert it into an atomic superposition state, and to restore it in its initial state of light. Faithful retrieval of the incoming signal relies on the elimination of spontaneous and stimulated emission. To this end, we propose a new protocol we have named « Revival Of Silenced Echo » (ROSE).

[pt] Neste trabalho, o Método dos Elementos Finitos é aplicado na análise eletromagnética e no projeto de cornetas coaxiais, utilizadas como alimentador em diversos tipos de antenas refletoras omnidirecionais. Estes alimentadores são usualmente excitados por conectores comercias, necessitando de uma estrutura de adaptação para conectar-se a corneta. Este adaptador é composto de seções não uniformes de guias de ondas coaxiais, sendo necessária à inclusão de anéis dielétricos para garantir a rigidez mecânica da antena. O correto dimensionamento desses anéis e das descontinuidades de guias de onda coaxiais que compõe a estrutura de alimentação permite estabelecer um compromisso entre minimização das perdas, alargamento da banda passante e rigidez mecânica. Devido a não uniformidade encontrada nestes dispositivos de microondas, métodos numéricos devem ser utilizados na predição do seu comportamento eletromagnético. Neste trabalho, o algoritmo baseado em Método de Elementos Finitos é utilizado para avaliar o desempenho detalhado dos diversos elementos que compõe o adaptador. Além disto, é comparado o desempenho de diversas formas e dimensões que caracterizam a corneta eletromagnética. Para validar os resultados do algoritmo, os resultados obtidos nas diversas etapas são comparados com aqueles obtidos via Método de Casamento de Modos. / [en] In this work, the Method of Finite Element is applied in the electromagnetic analysis and project of coaxial horns employed as feed in omnidirectional reflector antennas. These feeds are usually excited by standard commercial connectors that require an adaptor to connect to the coaxial horn. This adaptor is usually composed of non-uniform sections of coaxial wave guide, where dielectric rings are used to give the mechanical strength to the structure. The adequate dimensioning of these coaxial sections leads to a compromise between the minimization of return loss, the required frequency operation band, and the necessary mechanical rigidity of the structure. Due to the non-uniformities, numerical methods are required to predict the electromagnetic behavior. In this work, an algorithm based on Method of Finite Elements is used to evaluate the performance of the sections that composed the feed chain. The performance of different types of horn sections and dimensions are explored by using the algorithm. To validate the results obtained by using Method of Finite Elements they are compared with those obtained by employing Method of Mode Matching.

Made available in DSpace on 2016-07-27T14:21:40Z (GMT). No. of bitstreams: 1 Adriana da Silva Arantes Campos.pdf: 1130712 bytes, checksum: 034ea15f94e341fbfefc374b94bf4a0f (MD5) Previous issue date: 2009-01-06 / This study examined experientially behaviors considered mental processes by other psychological guidelines, but for Behavior Analysis, founded by the philosophy of Radical Behaviorism proposed by Skinner, are considered private verbal behaviors, as thought. To verify functional relations that artificial contingencies perform on participants performance and the relationship of these artificial contingencies with public and private pre-happenings in a mathematical task of troubleshooting. Two experiments were conducted with 6 participants in a subject design as his own control. Experiment I objected to verify functional relations of public pre-happenins and magnitude of strengthening. Experiment II aimed to verify if participants performance in a mathematical task of troubleshooting is function of artificial contingencies and the relationship of these contingencies with public and private pre-happenins during the task proposed. Results suggested that magnitudes of strengthening defined as points do not interfere in participants performance on solving mathematical equations with unknowns (private behavior), natural reinforcement seems to have controlled participants behavior in this type of task resolution. / O presente estudo analisou experimentalmente comportamentos considerados processos mentais por outras orientações psicológicas, mas que para a Análise do Comportamento, ciência fundamentada pela filosofia do Behaviorismo Radical proposta por Skinner, são considerados comportamentos verbais privados, tal como o pensamento. Com o objetivo de verificar as relações funcionais que contingências arbitrárias exercem sobre o desempenho de participantes e a relação dessas contingências arbitrárias com precorrentes públicos e privados em uma tarefa de solução de problemas matemáticos. Dois experimentos foram conduzidos com 6 participantes com um delineamento do sujeito como seu próprio controle. O Experimento I objetivou verificar as relações funcionais de precorrentes públicos e magnitude do reforço. O Experimento II teve como objetivo verificar se o desempenho dos participantes em uma tarefa de resolução de problemas matemáticos é função de contingências arbitrárias e a relação destas contingências com precorrentes públicos e privados na tarefa proposta. Os resultados sugeriram que magnitudes de reforço definidas como pontos não interferem no desempenho dos participantes em resolver equações matemáticas com incógnitas (comportamento privado), o reforço natural de resolver a tarefa parece ter controlado o comportamento dos participantes neste tipo de tarefa.

La présente thèse de doctorat prend pour objet d’étude la question des disparus de la fin de la guerre d’Algérie, en l’occurrence, à partir du cessez-le-feu du 19 mars jusqu’à la fin de l’année 1962. Ne pouvant restreindre notre investigation à cette seule période, nous avons élargi notre étude aux conséquences des disparitions. Cette démarche, qui a le mérite de suivre l’événement tragique de son apparition à sa prise en compte par les autorités et la société, avec son retentissement jusqu’à aujourd’hui se déroule en trois parties. La première partie « chercher les disparus » (chapitres 1 à 3), revient sur les recherches, par les autorités compétentes, de la personne physique ou de sa dépouille disparue d’une part, et présente d’autre part la relégation du fait de « chercher les disparus » en un objet de recherche scientifique. Les abus de langage liés à la polysémie du mot « disparu », la surenchère statistique et les usages politiques et mémoriels sont également mis en lumière afin de comprendre les raisons de cet angoissant mystère. La deuxième partie présente les modes opératoires adoptés, tels que les enlèvements et les arrestations arbitraires, afin de « faire disparaître » (chapitres 4 à 6). Les différents auteurs, cibles et mobiles de ces exactions sont ainsi examinés. Enfin, la troisième partie « vivre la disparition » (chapitres 7 à 9) revient essentiellement sur les répercussions économiques et psychologiques pour les familles et les proches confrontés, entre autres, à des problèmes pécuniaires, au poids des rumeurs et au deuil impossible lequel est lié à l’incertitude du sort des « ni morts, ni vivants ». / This doctoral thesis aims to study the issue of the people who went missing at the end of the Algerian War, namely from the cease-fire of 19 march until the end of 1962. As we were not able to restrict our investigation to this period alone, we expanded our study to the consequences of the disappearances. This approach, which follows the tragic event from its outset to its recognition by the authorities and society, including its impact to date, is in three parts. The first part, « searching for the disappeared » (chapters 1-3), revisits the search by the relevant authorities for the missing individual or their remains, and the relegation of the « search for the disappeared » to an object of scientific research. The misuse of language linked to the multiple meanings of the word « disappeared », statistical escalation and the political and memorial uses of the issue are also highlighted in order to understand the reasons behind this agonizing mystery. The second part presents the procedures used, such as abductions and arbitrary arrests, to « make people disappear » (chapters 4-6). The different perpetrators, targets and motives of these abuses are also examined. Finally, the third part « living with disappearance » (chapters 7-9) focuses largely on the economical and psychological repercussions for families and loved ones, which includes financial problems, rumors and the impossibility of mourning due to the uncertainty of the fate of « those who are neither dead nor alive ».

O fenômeno de molhamento, estudo de como um líquido se deposita em um sólido, apresenta problemas ainda em aberto, dos pontos de vista da modelagem física e da simulação numérica. O maior interesse acadêmico neste tipo de escoamento é a linha tríplice (ou linha de contato) formada da interação sólido-líquido-gás. A condição de contorno clássica de não escorregamento na interface líquido-sólido leva a uma singularidade no tensor de tensões nesta linha. Além disso, ainda não está estabelecido qual o melhor modelo para descrever o ângulo de contato formado entre a superfície livre e o substrato (o sólido). Neste trabalho, são discutidos métodos numéricos para a simulação de linhas de contato dinâmicas. Os efeitos da tensão superficial são estudados com a abordagem do princípio do trabalho virtual, o qual leva o problema à equações na formulação variacional, linguagem natural para o tratamento numérico com o método dos elementos finitos (FEM). O domínio é discretizado por uma malha não-estruturada de forma que as interfaces separadoras são explicitamente representadas pela malha. As derivadas temporais são tratadas em uma abordagem Lagrangeana-Euleriana arbitrária (ALE). Finalmente, são apresentados os resultados numéricos obtidos com o método ALE-FEM, discutindo alguns aspectos da sua convergência temporal e espacial. / Wetting phenomena, study of how of a liquid spreads out on a solid substrate, presents challenges both in physical modeling and in numerical simulation. The triple line (or contact line) formed by the solid-liquid-gas interaction has increasingly attracted the attention of the fluid dynamic community. The classical no-slip boundary condition on the liquid-solid interface leads to a singularity in the stress tensor at contact lines. Furthermore, there is no consensus on what the best model to describe the dynamics of the contact angle formed by the solid substrate and free surface. In this work, numerical methods for simulating dynamic contact lines are considered. The capillarity effects are studied in the approach of the virtual-work principle, which describes the problem in the variational formulation, natural language for numerical treatment with the finite element method (FEM). The domain is discretized by a dynamic unstructured mesh, where the separating interfaces are explicit represented by the mesh. Time derivatives present in the governing equations are treated with the arbitrary Lagrangian-Eulerian (ALE) framework. Finally, we discuss some temporal and spatial convergence issues ofthe ALE-FEM method.

<p>This thesis presents an approach to free-form surface manipulations, which conceptually improves an existing CAD system that constructs surfaces by smoothly interpolating a network of intersecting curves. There are no regularity requirements on the network, which already yields superior modelling capabilities compared to systems that are based on industry-standard NURBS surfaces.</p><p>Originally, the shape of such a surface can be modified only locally by manipulating a curve in the network. In this process there is an inherent danger that the curve is being pulled away from intersections that it has with other curves. When this happens, the network is invalidated as a surface representation, and many curves may have to be adjusted to restore network consistency and surface quality. This thesis contributes a method that solves these problems by propagating changes that are made in one curve to curves in its vicinity. How and to what extent curves react to changes is controlled by two parameters that can be varied along the curve that is being manipulated. Any curve may be constrained in one or more degrees of freedom. The integrity of the curve network is implicitly conserved, as well as the geometric continuity of the surface.</p><p>The result is a tool for the modification of curve-interpolating surfaces, which can easily be applied to large areas on models with any level of detail. This allows designers to concentrate on the creative process, rather than on planning chains of actions. They can explore different design variations, optimise shapes further, and generally be more productive.</p> / <p>Dette doktorgradsarbeidet presenterer en fremgangsmåte for formgivning og modifisering av datamaskinbaserte, skulpturerte flater. Metoden forbedrer et eksisterende system for data-assistert konstruksjon (DAK) som bygger dobbeltkrummede flater ved å interpolere et nettverk av skjærende kurver. Nettverket trenger ikke være regelmessig, noe som allerede gir bedre modelleringsmuligheter sammenliknet med systemer som er basert på standard NURBS flater.</p><p>En slik flate kan opprinnelig bare endres lokalt ved å dra i en kurve. I denne prosessen er det fare for at kurven blir dratt fra skjæringspunkter den har med andre kurver. Hvis dette skjer, representerer ikke nettverket en flate lenger, og mange kurver må justeres for å få tilbake integriteten i nettverket og kvaliteten i formen. Denne avhandlingen bidrar med en metode som løser disse problemene ved å spre endringer som blir gjort i en kurve til andre kurver i nærheten. Hvordan og i hvilken utstrekning kurvene reagerer på endringen styres av to parametre som kan varieres langs kurven som blir endret. Enhver kurve kan låses i en eller flere frihetsgrader. Integriteten til nettverket samt glattheten i formen blir bevart automatisk.</p><p>Resultatet er et redskap for modifikasjon av kurve-interpolerende flater som med letthet kan brukes på større områder av modeller med hvilken som helst grad av detalj. Dette gir designere muligheten til å konsentrere seg om det kreative, istedenfor å planlegge handlingsrekker. De kan utforske forskjellige designvariasjoner, optimalisere former ytterligere, og i det hele tatt være mer produktive.</p>

Semi-solid thixoforming processes rely on a material microstructure made of globular solid grains more or less connected to each other, thus developing a solid skeleton deforming into a liquid phase. During processing, the material structure changes with the processing history due to the agglomeration of the particles and the breaking of the grains bonds. This particular evolutive microstructure makes semi-solid materials behave as solids at rest and as liquids during shearing, which causes a decrease of the viscosity and of the resistance to deformation while shearing. Thixoforming of aluminum and magnesium alloys is state of the art and a growing number of serial production lines are in operation all over the world. But there are only few applications of semi-solid processing of higher melting point alloys such as steel. This can partly be attributed to the high forming temperature combined with the intense high temperature corrosion that requires new technical solutions. However the semi-solid forming of steels reveals high potential to reduce material as well as energy consumption compared to conventional process technologies, such as casting and forging. Simulation techniques exhibit a great potential to acquire a good understanding of the semi-solid material process. Therefore, this work deals with the development of an appropriate constitutive model for semi-solid thixoforming of steel. The constitutive law should be able to simulate the complex rheology of semi-solid materials, under both steady-state and transient conditions. For example, the peak of viscosity at start of a fast loading should be reproduced. The use of a finite yield stress is appropriate because a vertical billet does not collapse under its own weight unless the liquid fraction is too high. Furthermore, this choice along with a non-rigid solid formalism allows predicting the residual stresses after cooling down to room temperature. Several one-phase material modeling have been proposed and are compared. Thermo-mechanical modeling using a thermo-elasto-viscoplastic constitutive law has been developed. The basic idea is to extend the classical isotropic hardening and viscosity laws to the non solid state by considering two non-dimensional internal parameters. The first internal parameter is the liquid fraction and depends on the temperature only. The second one is a structural parameter that characterizes the degree of structural build up in the microstructure. Those internal parameters can depend on each other. The internal parameters act on the the viscosity law and on the yield surface evolution law. Different formulations of viscosity and hardening laws have been proposed and are compared to each other. In all cases, the semi-solid state is treated as a particular case, and the constitutive modeling remains valid over the whole range of temperature, starting from room temperature to above the liquidus. These models are tested and illustrated by mean of several representative numerical applications.