Prouver l’équivalence de programmes écrits dans un langage fonctionnel avec références est un problème notoirement difficile. L’objectif de cette thèse est de proposer un système logique dans lequel de telles preuves peuvent être formalisées, et dans certains cas inférées automatiquement. Dans la première partie, une méthode générique d’extension de la théorie des types dépendants est proposée, basée sur une interprétation du forcing vu comme une traduction de préfaisceaux de la théorie des types. Cette extension dote la théorie des types de constructions récursives gardées, qui sont utilisées ensuite pour raisonner sur les références d’ordre supérieure. Dans une deuxième partie, nous définissons une sémantique des jeux nominale opérationnelle pour un langage avec références d’ordre supérieur. Elle marie la structure catégorique de la sémantique des jeux avec une représentation sous forme de traces de la dénotation des programmes, qui se calcule de manière opérationnelle et dispose donc de bonnes propriétés de modularité. Cette sémantique nous permet ensuite de prouver la complétude de relations logiques à la Kripke définit de manière directe, via l’utilisation de types récursifs gardés, sans utilisation de la biorthogonalité. Une telle définition directe nécessite l’utilisation de mondes omniscient et un contrôle fin des locations divulguées. Finalement, nous introduisons une logique temporelle qui donne un cadre pour définir ces relations logiques à la Kripke. Nous ramenons alors le problème de l’équivalence contextuelle à la satisfiabilité d’une formule de cette logique générée automatique, c’est à dire à l’existence d’un monde validant cette formule. Sous certaines conditions, cette satisfiabilité peut être décidée via l’utilisation d’un solveur SMT. La complétude de notre méthode devrait permettre d’obtenir des résultats de décidabilité pour l’équivalence contextuelle de certains fragment du langage considéré, en fournissant un algorithme pour construire de tels mondes. / Proving program equivalence for a functional language with references is a notoriously difficult problem. The goal of this thesis is to propose a logical system in which such proofs can be formalized, and in some cases inferred automatically. In the first part, a generic extension method of dependent type theory is proposed, based on a forcing interpretation seen as a presheaf translation of type theory. This extension equips type theory with guarded recursive constructions, which are subsequently used to reason on higher-order references. In the second part, we define a nominal game semantics for a language with higher-order references. It marries the categorical structure of game semantics with a trace representation of denotations of programs, which can be computed operationally and thus have good modularity properties. Using this semantics, we can prove the completeness of Kripke logical relations defined in a direct way, using guarded recursive types, without using biorthogonality. Such a direct definition requires omniscient worlds and a fine control of disclosed locations. Finally, we introduce a temporal logic which gives a framework to define these Kripke logical relations. The problem of contextual equivalence is then reduced to the satisfiability of an automatically generated formula defined in this logic, i.e. to the existence of a world validating this formula. Under some conditions, this satisfiability can be decided using a SMT solver. Completeness of our methods opens the possibility of getting decidability results of contextual equivalence for some fragments of the language, by giving an algorithm to build such worlds.

Le travail abordé dans le cadre de cette thèse constitue une contribution à la compréhension des mécanismes d'endommagement et à l'identification d'un critère de durée de vie du polyamide 6,6 renforcé à 30% en masse de fibres de verre courtes (PA66GF30), obtenu par moulage par injection, en prenant en compte les effets de la température, de la teneur en eau et des orientations de fibres induites par le procédé de fabrication. Ainsi, dans un premier temps, une campagne d'essais de caractérisation du comportement mécanique du PA66GF30, pour différentes orientations de fibres, températures, teneurs en eau et vitesses de sollicitation, a été menée. Ceux-ci ont mis en évidence le rôle de la matrice polyamide 6,6 (PA66) dans la dépendance à la température, à la teneur en eau et à la vitesse de sollicitation du comportement du PA66GF30. Dans ce sens, une triple équivalence Temps-Température-Teneur en eau du comportement du PA66 a été démontrée à l'aide de courbes maîtresses obtenues par DMA. Une nouvelle expression de la vitesse équivalente à une température et une teneur en eau de référence en a alors été déduite, dans l'objectif de pouvoir rendre compte de cette triple sensibilité dans un futur modèle de comportement. Ensuite, une campagne d'essais de fatigue dans des conditions environnementales contrôlées, a permis d'étudier les effets de l'orientation des fibres, de la température et de la teneur en eau, sur la tenue en fatigue d'éprouvettes en PA66GF30. L'analyse mécanique de ces essais a abouti à l'identification d'un critère de durée de vie, tenant compte de l'ensemble des facteurs influents. Enfin, la microtomographie des rayons X sous faisceau synchrotron a été utilisée pour étudier les mécanismes d'endommagement en fatigue présents dans le PA66GF30, suite à des analyses post-mortem et des expériences in-situ. Le dépouillement de ces observations a permis de faire la lumière sur les mécanismes d'endommagement, et de proposer un nouveau scénario d'endommagement en fatigue. / The work discussed in this thesis is a contribution to the understanding of damage mechanisms and identifying a fatigue criterion of polyamide 6,6 reinforced with 30 wt% of short glass-fibres (PA66GF30), obtained by injection molding, taking into account the effects of temperature, water content and fiber orientations induced by the process. Thus, in a first step, the characterization of the mechanical behavior of PA66GF30 for different fiber orientations, temperatures, moisture contents and strain rates, was conducted. They have highlighted the role of the polyamide matrix (PA66) in the temperature, water content and strain rate behavior dependence of PA66GF30. In this sense, Time-Temperature-Water content equivalence of mechanical behaviour of PA66 was demonstrated using master curves obtained by DMA. A new expression of the equivalent strain rate at temperature reference and a water content reference was then deduced, and it is now possible to relay the triple sensitivity to strain rate, temperature and water content in only one term. Then, a fatigue test campaign in controlled environmental conditions, allowed to study the effects of fiber orientation, temperature and water content on the fatigue behaviour of PA66GF30 specimens. The mechanical analysis of these tests has led to the identification of a fatigue life criterion, taking into account all the influencing factors. Finally, the X-ray microtomography under synchrotron beam was used to study the mechanisms in fatigue damage present in the PA66GF30, following post-mortem analysis and in-situ experiments. The analysis of these observations has shed light on the mechanisms damage and, it has been proposed a new fatigue damage scenario.

Neste trabalho estudamos algumas propriedades do espaço das funções contínuas munido da topologia da convergência pontual. Começamos estudando o espaço Cp(X) de forma geral, verificando que propriedades topológicas principais valem em Cp(X), usando teoremas de dualidade entre X e Cp(X). Em seguida estudamos a relação da estrutura topológica de X e a estrutura algébrica e topológica de Cp(X), onde o Teorema de Nagata é fundamental. Observamos algumas propriedades de X que são preservadas por l-equivalência ou t-equivalência, ou seja, que são determinadas pela estrutura linear topológica, ou pela estrutura topológica de Cp(X), respectivamente. Por último estudamos as condições para que Cp(X) seja um espaço de Lindelöf. Concluímos com a prova de Okunev de que o número de Lindelöf de Cp(X) é igual ao número de Lindelöf de Cp(X)xCp(X), para espaços fortemente zero-dimensionais X. / In this work we study some properties of the space of continuous functions endowed with the topology of pointwise convergence. We begin by studying the space Cp(X) in general terms, verifying that the main topological properties are valid in Cp(X), using duality theorems between X and Cp(X). Next we study the relationship between the topological structure of X and the algebraic as well as topological structure of Cp(X), in which the Nagata theorem theorem is essential. We observe some properties of X, which are preserved by l-equivalence or t-equivalence, i.e., which are respectively determined either by the linear topological structure of Cp(X) or by its topological one. Finally we study in which conditions Cp(X) is a Lindelöf space. We conclude with the proof of Okunev that the Lindelöf number of Cp(X) is equal to the Lindelöf number of Cp(X)xCp(X), for strongly zero-dimensional spaces X.

Made available in DSpace on 2016-07-27T14:18:43Z (GMT). No. of bitstreams: 1 Antonio Carlos Godinho dos Santos.pdf: 1568718 bytes, checksum: c84bef3972697df980163958f88b2a62 (MD5) Previous issue date: 2010-12-17 / Literature points that children in sixth grade of Middle School show difficulty to solve fractional problems that demand the concept of proportion. Behavior analysis affords efficient techniques and procedures to form stimulus classes that originate concepts. Nevertheless, few production about concepts formation based on this perspective were achieved on Math field. The present study s goal is to investigate the effect of conditional relationship training between: fractional pictorial stimulus (group A) and numerical (group B) and, numerical (group B) and numerical (group C), both performed by means of the stimulus equivalence paradigm, concerning formation of equivalence classes with fractional stimulus, generalization of responses to new situations and the expansion of the formed classes after the training among members of group A and a new group of numerical stimulus (D). Three experiments were conducted with an amount of 90 children enrolled in State educational system in Goiania and Trindade cities, falling into both genres whose ages range from 11 to 14 years old. The participants, divided in experimental and control groups, were initially submitted to a pencil and paper evaluation about fractional problems (IE) expressed in absolute and relative way. In the first experiment, one group was exposed only to the training with programmed instructions about fractional problem resolutions (PIG), another one only to conditional relationship training (EQG) and a third one to both trainings (EQPIG). In experiment 2, one group handled concrete material presented in form of fraction (CMG), another one went through the training of the two conditions (CMGEQ). In experiment 3, one group trained conditional relationships with fractional stimulus (EQG), another one went through the training of compounding (copying) fractions from models visually presented and conditional relationship tests (EQCTG) and to a third group was asked the exposition of the strategies adopted to solve problems during the trainings and conditional relationship tests (GEQR). At the end of the procedures all of them went through a final evaluation (FE) identical to the initial evaluation (IE). The control participants took only the initial and final evaluations. The results show that participants presented difficulty to solve problems expressed in relative way. Generally, all the participants of the experimental groups of the three experiments formed three fractional stimulus classes (ABC), however there was no expansion of these classes after the training between one of its members and a new stimulus. The highest generalization indexes occurred in the first programmed test and decreased in the second evaluation. Only participants of experimental groups presented systematic gains from IE to FE, considering that the highest performances to problems with fractions occurred to participants in EQPIG group. / A literatura aponta que crianças de final do sexto ano do ensino fundamental demonstram dificuldade para resolver problemas fracionários que exigem o conceito de proporção. A análise do comportamento dispõe de técnicas e procedimentos eficientes para formar classes de estímulos que dão origem a conceitos. Contudo, poucos trabalhos sobre formação de conceitos embasados nessa perspectiva foram realizados no campo da matemática. O objetivo do presente estudo foi investigar o efeito do treino de relações condicionais entre: estímulos fracionários pictóricos (conjunto A) e numéricos (conjunto B) e, entre estímulos fracionários numéricos (conjunto B) e numéricos (conjunto C), ambos realizados por meio do procedimento de pareamento com o modelo a partir do paradigma de equivalência de estímulos sobre a formação de classes de equivalência com estímulos fracionários, tendo sido avaliada também a generalização da resposta a novas situações e a expansão das classes formadas após o treino entre os membros do conjunto A e um novo conjunto de estímulos numéricos (D). Foram conduzidos três experimentos com um total de 90 crianças matriculadas na rede estadual de ensino das cidades de Goiânia e Trindade, de ambos os sexos cujas idades variaram entre 11 e 14 anos. Os participantes, divididos em grupos experimentais e de controle, passaram inicialmente por uma avaliação do tipo lápis e papel sobre problemas fracionários (AI) expressos de forma absoluta e relativa. No primeiro experimento, um grupo foi exposto apenas ao treino com instruções programadas sobre a resolução de problemas fracionários (GIP), outro apenas ao treino de relações condicionais (GEQ) e um terceiro a ambos os treinos (GIPEQ). No experimento 2, um grupo manuseou material concreto apresentado em forma de fração (GMC), outro fez treino de relações condicionais entre estímulos fracionários (GEQ) e um terceiro passou pelo treino das duas condições (GMCEQ). No experimento 3, um grupo fez treino de relações condicionais com estímulos fracionários (GEQ), outro passou pelo treino de composição (cópia) de frações a partir de modelos apresentados visualmente mais o treino de relações condicionais (GEQTC) e a um terceiro grupo foi solicitado o relato das estratégias que eles adotaram para resolver os problemas durante os treinos e testes das relações condicionais(GEQR). Ao final dos procedimentos todos eles passaram por uma avaliação final (AF) idêntica à avaliação inicial (AI) aplicada no início do estudo. Os participantes controles fizeram apenas as avaliações inicial e final. Os resultados mostraram que os participantes demonstraram dificuldade para resolver problemas expressos na forma relativa. Em geral, todos os participantes dos grupos experimentais dos três experimentos formaram três classes de estímulos fracionários (ABC), mas não houve expansão dessas classes após o treino entre um de seus membros e um novo estímulo. Os maiores índices de generalização ocorreram no primeiro teste programado e decaíram na segunda avaliação. Apenas os participantes dos grupos experimentais apresentaram ganhos sistemáticos da AI para a AF, sendo que as maiores porcentagens de acertos para problemas com frações expressas de forma relativa ocorreram para os participantes do grupo GIPEQ.

Neste trabalho estudamos algumas propriedades do espaço das funções contínuas munido da topologia da convergência pontual. Começamos estudando o espaço Cp(X) de forma geral, verificando que propriedades topológicas principais valem em Cp(X), usando teoremas de dualidade entre X e Cp(X). Em seguida estudamos a relação da estrutura topológica de X e a estrutura algébrica e topológica de Cp(X), onde o Teorema de Nagata é fundamental. Observamos algumas propriedades de X que são preservadas por l-equivalência ou t-equivalência, ou seja, que são determinadas pela estrutura linear topológica, ou pela estrutura topológica de Cp(X), respectivamente. Por último estudamos as condições para que Cp(X) seja um espaço de Lindelöf. Concluímos com a prova de Okunev de que o número de Lindelöf de Cp(X) é igual ao número de Lindelöf de Cp(X)xCp(X), para espaços fortemente zero-dimensionais X. / In this work we study some properties of the space of continuous functions endowed with the topology of pointwise convergence. We begin by studying the space Cp(X) in general terms, verifying that the main topological properties are valid in Cp(X), using duality theorems between X and Cp(X). Next we study the relationship between the topological structure of X and the algebraic as well as topological structure of Cp(X), in which the Nagata theorem theorem is essential. We observe some properties of X, which are preserved by l-equivalence or t-equivalence, i.e., which are respectively determined either by the linear topological structure of Cp(X) or by its topological one. Finally we study in which conditions Cp(X) is a Lindelöf space. We conclude with the proof of Okunev that the Lindelöf number of Cp(X) is equal to the Lindelöf number of Cp(X)xCp(X), for strongly zero-dimensional spaces X.

Les travaux présentés dans ce mémoire s'inscrivent dans l'étude et l'amélioration des modèles d'endommagement continus régularisés (non locaux), l'objectif étant d'étudier la transition entre un champ d'endommagement continu défini sur l'ensemble d'une structure et un modèle discontinu de fissuration macroscopique.La première étape consiste en l'étude semi-analytique d'un problème unidimensionnel (barre en traction) visant à identifier une famille de lois d'interface permettant de basculer d'une solution non homogène obtenue avec un modèle continu à gradient d'endommagement vers un modèle discontinu de fissuration cohésive. Ce passage continu / discontinu est construit de telle sorte que l'équivalence énergétique entre les deux modèles soit assurée, et reste exacte quelque soit le niveau de dégradation atteint par le matériau au moment où cette transition est déclenchée.Cette stratégie est ensuite étendue au cadre 2D (et 3D) éléments finis dans le cas de la propagation de fissures rectilignes (et planes) en mode I. Une approche explicite basée sur un critère de dépassement d'une valeur « critique » de l'endommagement est proposée afin de coupler les modèles continus et discontinus au sein d'un même calcul quasi-statique par éléments finis. Enfin, plusieurs résultats de simulations menées avec cette approche couplée sont présentés. / The present work deals with the study and the improvement of regularized (non local) damage models. It aims to study the transition from a continuous damage field distributed on a structure to a discontinuous macroscopic failure model.First, an analytical one-dimensional study is carried out (on a bar submitted to tensile loading) in order to identify a set of interface laws that enable to switch from an inhomogeneous solution obtained with a continuous gradient damage model to a cohesive zone model. This continuous / discontinuous transition is constructed so that the energetic equivalence between both models remains ensured whatever the damage level reached when switching.This strategy is then extended to the bi-dimensional (and tri-dimensional) case of rectilinear (and plane) crack propagation under mode I loading conditions, in a finite element framework. An explicit approach based on a critical damage criterion that allows coupling both continuous and discontinuous approaches is then proposed. Finally, results of several simulations led with this coupled approach are presented.

Le problème d’´équivalence de code joue un rôle important dans la théorie de code et la cryptographie basée sur le code. Cela est dû à son importance dans la classification des codes ainsi que dans la construction et la cryptanalyse des cryptosystèmes à base de codes. Il est également lié à un problème ouvert d’isomorphisme de graphes, un problème bien connu dans le domaine de la théorie de la complexité. Nous prouvons pour les codes ayant un hull trivial qu’il existe une réduction polynomiale de l’équivalence par permutation de codes à l’isomorphisme de graphes. Cela montre que cette sous-classe d’équivalence de permutation n’est pas plus dure que l’isomorphisme de graphes. Nous introduisons une nouvelle méthode pour résoudre le problème d’équivalence de code. Nous développons des approches algébriques pour résoudre le problème dans ses deux versions : en permutation et en diagonale. Nous construisons un système algébrique en établissant des relations entre les matrices génératrices et les matrices de parité des codes équivalents. Nous nous retrouvons avecun système plusieurs variables d’équations linéaires et quadratiques qui peut être résolu en utilisant des outils algébriques tels que les bases de Groebner et les techniques associées. Il est possible en théorie de résoudre l’équivalence de code avec des techniques utilisant des bases de Groebner. Cependant, le calcul en pratique devient complexe à mesure que la longueur du code augmente. Nous avons introduit plusieurs améliorations telles que la linéarisation par bloc et l’action de Frobenius. En utilisant ces techniques, nous identifions de nombreux cas où le problème d’équivalence de permutation peut être résolu efficacement. Notre méthode d’équivalence diagonale résout efficacement le problème dans les corps de petites tailles, à savoir F3 et F4. L’augmentation de la taille du corps entraîne une augmentation du nombre de variables dans notre système algébrique, ce qui le rend difficile à résoudre. Nous nous intéressons enfin au problème d’isomorphisme de graphes en considérant un système algébrique quadratique pour l’isomorphisme de graphes. Pour des instances tirées aléatoirement, le système possède des propriétés intéressantes en termes de rang de la partie linéaire et du nombre de variables. Nousrésolvons efficacement le problème d’isomorphisme de graphes pour des graphes aléatoires avec un grand nombre de sommets, et également pour certains graphes réguliers tels que ceux de Petersen, Cubical et Wagner.123 / Code equivalence problem plays an important role in coding theory and code based cryptography.That is due to its significance in classification of codes and also construction and cryptanalysis of code based cryptosystems. It is also related to the long standing problem of graph isomorphism, a well-known problem in the world of complexity theory. We introduce new method for solving code equivalence problem. We develop algebraic approaches to solve the problem in its permutation and diagonal versions. We build algebraic system by establishing relations between generator matrices and parity check matrices of the equivalent codes. We end up with system of multivariables of linear and quadratic equations which can be solved using algebraic tools such as Groebner basis and related techniques. By using Groebner basis techniques we can solve the code equivalence but the computation becomes complex as the length of the code increases. We introduced several improvements such as block linearization and Frobenius action. Using these techniques we identify many cases where permutation equivalence problem can be solved efficiently. Our method for diagonal equivalence solves the problem efficiently in small fields, namely F3 and F4. The increase in the field size results in an increase in the number of variables in our algebraic system which makes it difficult to solve. We introduce a new reduction from permutation code equivalence when the hull is trivial to graph isomorphism. This shows that this subclass of permutation equivalence is not harder than graph isomorphism.Using this reduction we obtain an algebraic system for graph isomorphism with interesting properties in terms of the rank of the linear part and the number of variables. We solve the graph isomorphism problem efficiently for random graphs with large number of vertices and also for some regular graphs such as Petersen, Cubical and Wagner Graphs.

In this thesis we consider various questions and problems about graphs that appear in the framework of descriptive set theory. The main object of study are graphons, graphings and variations of the graph G0. We establish an approach to the compactness of the graphon space via the weak* topology and introduce the notion of a fractional isomorphism for graphons. We use a variant of the G0-dichotomy in the context of the classification problem. Finally, we show a measurable version of the Vizing's theorem for graphings. 1

No description available.

Thesis (M.A. (Translation Studies and Linguistics)) -- University of Limpopo, 2008 / Refer to document