Diagonally dominant matrices arise in many applications. In this work, we exploit the structure of diagonally dominant matrices to provide sharp entrywise relative perturbation bounds. We first generalize the results of Dopico and Koev to provide relative perturbation bounds for the LDU factorization with a well conditioned L factor. We then establish relative perturbation bounds for the inverse that are entrywise and independent of the condition number. This allows us to also present relative perturbation bounds for the linear system Ax=b that are independent of the condition number. Lastly, we continue the work of Ye to provide relative perturbation bounds for the eigenvalues of symmetric indefinite matrices and non-symmetric matrices.

This thesis is titled The influence of a nonlinear Power Take Off on a Wave Energy Converter. It looks at the effect that having a nonlinear Power Take Off (PTO) has on an inertial referenced, slack moored, point absorber, Wave Energy Converter (WEC). The generic device studied utilizes relative heave motion between an axi-symmetric cylinder and an internal mass, for the PTO to operate between. The PTO is the part of the WEC that transforms the relative motion into electricity. In this work, three different types of nonlinear PTO and a linear PTO are presented, tested, analysed and compared. The three nonlinear PTO types are: • A PTO that extracts energy in only one direction, either in relative compression or expansion. • A linear PTO and an additional endstop or peripheral PTO, that can only extract energy when the relative position of the internal mass has reached a pre-determined position. • A PTO that has damping forces that are quadratically proportional to the relative velocity. A numerical simulation has been built based upon a Runge-Kutta time series progression. The model uses the summation of the excitation force from the waves, the radiation force from the movement of the cylinder, the buoyancy force and the PTO forces. These combine to cause acceleration of the mass of the external cylinder, with an equal and opposite PTO force acting on the internal mass. The excitation force and added mass values are obtained from the boundary element method software, WAMIT. Prony’s method is used to obtain an approximate radiation force, based upon the radiation time force history. This numerical model operates on both a 1:40 scale and a full sized model. The numerical model finds the optimal PTO parameters, for different PTO setups, in irregular sea states. This optimum is based on the power extracted as well as indications of the reliability and lifetime of the system. The numerical simulation presents results showing how the nonlinearity of the PTO influences the motions of the WEC, resulting in dissimilarities between the Response Amplitude Operator (RAO) results, obtained from regular seas, and the Linear Transfer Function (LTF), found from irregular sea testing. The experimental model has been tested in the Curved Wave Tank facility at the University of Edinburgh, with a 1:40 scaled model. It used a central rod both as a support structure and to limit the movement of the cylinder and internal mass to heave. Between the cylinder and internal mass a spring and pneumatic damper operate in parallel, in various setups. It was tested in regular and irregular sea states and the position of the internal mass and cylinder was monitored. The experimental model was tested to ascertain the time series motions, RAO, LTF, the relative phase between the bodies and the power extracted for different wave climates. The numerical and experimental work were compared to allow confidence in both models. They showed relatively good agreement for the RAOs, LTFs and predictions of the relative phase but there was discrepancies in the predicted power for both regular and irregular seas. This difference is due to the difficulties in obtaining the relative velocities in the experimental model, resulting in a significant error in power prediction, since the power is proportional to the square of the relative velocities. The conclusions show that having a mono-directional PTO as opposed to a bi-directional PTO results in an approximately equal or greater power extraction in a variety of different sea states. An additional endstop or peripheral damper can increase the total power that a WEC extracts, in some situations, and may be advantageous depending upon the other potential benefits it brings to the WEC.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal.

Information theoretical measures are used to design, from first principles, an objective function that can drive a learning machine process to a solution that is robust to perturbations in parameters. Full analytic derivations are given and tested with computational examples showing that indeed the procedure is successful. The final solution, implemented by a robust learning machine, expresses a balance between Shannon differential entropy and Fisher information. This is also surprising in being an analytical relation, given the purely numerical operations of the learning machine.

Syfte Syftet med studien var att undersöka om relative age effect förekommer på svenska längdskidgymnasier bland antagna elever och hur eventuell förekomst ser ut bland flickor och pojkar samt hur det ser ut vid Riksidrottsgymnasierna och de Nationell Idrottsutbildningarna. Metod För att besvara syftet har statistisk data analyserats med hjälp av ett Pearson Chi2-test. Analysen har skett genom att undersöka födelsemånad, kön och inriktning bland antagna elever vid svenska längdskidgymnasier mellan åren 2010-2015. Resultat Resultatet visar att det finns signifikanta bevis på att relative age effect förkommer på svenska längdskidgymnasier (p= <0,05). Vid uppdelning av respektive utbildningsinriktning kunde det endast konstateras förekomst av relative age effect vid Riksidrottsgymnasierna, vid uppdelning av kön går det inte att säkerställa en signifikant förekomst av relative age effect. Slutsatser Studiens resultat visar att det förekommer relative age effect på svenska längdskidgymnasier, dock varierar den mellan de analyserade undervariablerna. Tydligast förekommer relative age effect vid Riksidrottsgymnasierna, medan det vid Nationell Idottsutbildning, kunde det inte konstateras bevis för en signifikant överrepresentation av elever födda under första kvartalet. När flickor och pojkar analyserades separat kunde ingen relative age effect konstateras. Det behövs mer forskning inom området för att få en överblick av konsekvenserna av relative age effect på längdskidgymnasier men även konsekvenserna för svensk längdskidåkning i stort. / Aim The purpose of this study was to investigate if relative age effect occurs at Swedish cross-country ski schools and if there were a differences in the presence among boys and girls, and between Riksidrottsgymnasier (Country-wide sport academy) and Nationell Idrottsutbildning (National sport academy). Method To answer the aim, statistical data were analyzed using a Pearson Chi2-test. The analysis has been done by examining birth month, sex and education among students at Swedish cross-country ski schools between 2010-2015. Results The result shows that there is a significant evidence of relative age effect at Swedish cross-country ski schools (p= <0,05). When the results were partial between Riksidrottsgymnasier and Nationell Idrottsutbildning, the relative age effect was only seen at Riksidrottsgymnasierna. There were no significant evidence for relative age effect when boys and girls where analyzed separately. Conclusion The results show that relative age effect appears on Swedish cross-country ski schools; however, there were some varies between analyzed variables. Strongest evidence for relative age effect was seen at Riksidrottsgymnaiserna. Further research is needed to get a overview of the consequences of relative age effect at cross-country ski schools, and the consequences it has in Swedish cross-country skiing overall, both for younger and older skiers.

Le béton est un matériau hétérogène complexe dont les déformations comportent une partie différée qui est affectée par un grand nombre de facteurs tels que la température, l'humidité relative et l'évolution de la microstructure. La prise en compte des déformations différées et en particulier du fluage est indispensable dans le calcul des ouvrages en béton tels que ceux destinés à stocker des déchets radioactifs. Ce travail de thèse a pour objectifs : (1) de développer un modèle de fluage simple et robuste pour le béton en faisant appel à la micromécanique et en tenant compte de l'endommagement et des effets thermiques et hydriques ; (2) d'implanter numériquement le modèle développé dans un code de calcul par éléments finis de façon à pouvoir simuler le comportement d'éléments de structure simples en béton. Pour atteindre ce double objectif, le travail est scindé en trois parties. Dans la première partie, le matériau cimentaire est à l'échelle microscopique supposé être constitué d'une matrice viscoélastique linéaire caractérisée par un modèle de Maxwell généralisé et de phases particulaires représentant les granulats élastiques et les pores. Le schéma micromécanique de Mori-Tanaka, la transformée de Laplace-Carson et son inversion sont alors utilisés pour obtenir dans l'espace temporel des estimations analytiques ou numériques de ses paramètres mécaniques et hydromécaniques. Ensuite, le modèle micromécanique de fluage obtenu est couplé au modèle d'endommagement de Mazars via le concept de pseudo-déformations introduit par Schapery. Les paramètres intervenant dans le modèle viscoélastique endommageable ainsi établi sont systématiquement identifiés à l'aide de données expérimentales. Enfin, la prise en compte des effets de la température et de l'humidité relative dans le modèle viscoélastique endommageable est basée sur la méthode du temps équivalent ; l'efficacité de cette approche est démontrée et discutée dans le cas de chargements simples de fluage / Concrete is a complex heterogeneous material whose deformations include a delayed part that is affected by a number of factors such as temperature, relative humidity and microstructure evolution. Taking into account differed deformations and in particular creep is essential in the computation of concrete structures such as those dedicated to radioactive waste storage. The present work aims: (1) at elaborating a simple and robust model of creep for concrete by using micromechanics and accounting for the effects of damage, temperature and relative humidity; (2) at numerically implementing the creep model developed in a finite element code so as to simulate the behavior of simple structural elements in concrete. To achieve this twofold objective, the present work is partitioned into three parts. In the first part the cement-based material at the microscopic scale is taken to consist of a linear viscoelastic matrix characterized by a generalized Maxwell model and of particulate phases representing elastic aggregates and pores. The Mori-Tanaka micromechanical scheme, the Laplace-Carson transform and its inversion are then used to obtain analytical or numerical estimates for the mechanical and hydromechanical parameters of the material. Next, the original micromechanical model of creep is coupled to the damage model of Mazars through the concept of pseudo-deformations introduced by Schapery. The parameters involved in the creep-damage model thus established are systematically identified using available experimental data. Finally, the effects of temperature and relative humidity are accounted for in the creep-damage model by using the equivalent time method; the efficiency of this approach is demonstrated and discussed in the case of simple creep tests

Essa dissertação pretende esclarecer o estatuto das orações relativas do karitiana (tupi- Arikém). Orações relativas podem ser classificadas como relativas de núcleo externo (RNE) e de núcleo interno (RNI), sendo o principal critério de diferenciação a posição do núcleo em relação à relativa: relativas com o núcleo adjacente à oração subordinada são classificadas como RNEs, enquanto aquelas com o núcleo interno à subordinada são RNIs (DE VRIES, 2006, CULY, 1990). Outro critério utilizado é a marcação de caso: se o núcleo estiver marcado com o caso exigido pelo verbo da matriz, a relativa será uma RNE; se for aquele exigido pelo verbo da relativa, ela será uma RNI (COLE, 1987). Dentro desse quadro, as orações do karitiana são difíceis de classificar: por um lado, o núcleo aparece sempre deslocado para a periferia esquerda (STORTO, 1999), algo característico de RNEs; por outro, a marcação de caso no núcleo segue o padrão de RNIs. À luz do trabalho de Basilico (1996), hipotetizamos que as relativas do karitiana seriam RNIs com frontalização opcional do núcleo. Se for o caso, é esperado que relativas com núcleos não frontalizados sejam permitidas. Montamos então um experimento para verificar se as relativas poderiam ter seus núcleos não frontalizados e testamos 14 falantes com uma metodologia de produção elicitada. Os resultados mostram que, embora haja uma tendência pela frontalização, núcleos não frontalizados são permitidos na língua, pois há casos de relativas de sujeito com a ordem OSV e de relativas de objeto SOtiV, OSV e SOV. Também foram produzidas relativas de objeto sem o morfema de foco do objeto , indicando que ele não é imprescindível para a relativização. Esse quadro aproxima nossas relativas das RNIs, pois RNEs não podem ter núcleos em outras posições além da periferia da oração relativa. Também analisamos propostas de análise sintática para as diversas ordens de palavras coletadas em nosso experimento. Vemos que aquelas que assumem algum tipo de deslocamento do núcleo para Spec de uma projeção de periferia seja CP ou AspP incorrem em diversos problemas, como a impossibilidade de derivar relativas de objeto SOtiV e a incapacidade de excluir estruturas agramaticais com advérbios. Assim, propomos que a frontalização do núcleo é uma adjunção a AspP. Dados de orações relativas com advérbios nos levam ainda a postular que, em relativas de objeto, a frontalização do núcleo ocorre em duas etapas: primeiro um movimento para Spec, vP e, em seguida, a frontalização para uma posição de adjunto de AspP. Essa primeira etapa do movimento seria marcada pela presença de em v e estaria na base do sincretismo desse morfema, que também está presente em perguntas qu- de objeto e em construções de foco do objeto. Por fim, oferecemos ainda uma análise da correlação entre a presença do morfema e a frontalização do objeto a partir do modelo de fases de Chomsky (2000, 2001), admitindo que o movimento do objeto para a borda do sintagma verbal seria uma operação sintática imprescindível para a subsequente frontalização do objeto / This dissertation aims to clarify the status of relative clauses in Karitiana (Tupi-Arikém). Relative clauses are traditionally classified as externally (EHRC) or internally-headed (IHRC) and the main criterion for their differentiation is the heads position: EHRCs have their heads adjacent to the relative clause itself, whereas IHRCs have internal heads (DE VRIES, 2006, CULY, 1990). Another criterion is case-marking: if the head is marked with the case demanded by the matrix verb, the relative is an EHRC; if it is the one demanded by the embedded verb, it will be an IHRC (COLE, 1987). Within this framework, karitiana relative clauses are hard to classify: on one hand, the head always appears fronted to the left periphery (STORTO, 1999), which resembles the pattern found in EHRCS; on the other hand, the case-marking on the head is similar to IHRCs. In the light of Basilicos (1996) work, one can hypothesize that karitiana relative clauses are IHRCs with optional head frontalization. In this case, it is expected that relatives with non-fronted heads will be allowed in the language. An experiment was designed in order to verify if karitiana relatives could have non-fronted heads and 14 speakers were tested with an elicited production methodology. The results show that, although there is a preference for frontalization, non-fronted heads are possible in the language, since subject relatives OSV and object relatives SOtiV, OSV and SOV were produced. There are also cases of object relatives without the object focus morpheme , indicating that it is not indispensable in relative clause formation. These results bring karitiana relatives closer to IHRCs, because EHRCs cannot have their heads in any other positions than in the periphery of the clause. We also discuss some syntactic proposals for the word orders found in our experiment, claiming that those which assume head dislocation to Spec of CP and AspP face some problems, such as the derivation of SOtiV object relatives and ungrammatical structures with adverbs. Therefore, our proposal is that the frontalization of the head is an adjunction to AspP. Paradigms of relative clauses with adverbs also show that, in object relatives, the frontalization of the head occurs in two steps: first the head moves to Spec, vP and then it is further fronted to the position of adjunct of Spec, AspP. The first step is marked with on v and it underlies the syncretism of this morpheme, which is also present in object wh- questions and object focus constructions. Finally, the correlation between and the frontalization of the head is analyzed within the phase theory framework (CHOMSKY, 2000, 2001) and it is assumed that object movement to vPs edge is a syntactic requirement for further frontalization

Mudanças na taxa de acumulação e na natureza da matéria orgânica sedimentar, e suas relações com as variações climáticas e flutuações do nível relativo do mar durante os últimos 8.500 anos cal. A.P. foram investigadas em dois testemunhos provenientes do Saco de Ribeira, Enseada do Flamengo, e junto à praia do Lázaro, Enseada da Fortaleza, na região de Ubatuba, litoral norte do Estado de São Paulo. Para o estudo as amostras coletadas foram submetidas a diversas análises - granulometria, teor de CaCO3, teores de Carbono orgânico, Nitrogênio e Enxofre totais, e razões isotópicas de carbono. Datações 14C, calibradas para a idade calendário, permitiram estabelecer um modelo de idades para as variações sedimentares detectadas. Intervalos de clima mais quente, 8.000-6.000 anos cal. A.P., foram acompanhados de um aumento na taxa de acumulação de carbono T.A.C. (10-13 g m-2 ano-1), seguida de diminuição após 5.000 anos cal. A.P., para valores inferiores a 2 g m-2 ano-1. Variações climáticas regionais, possivelmente relacionadas a chuvas torrenciais, próximos ao máximo transgressivo, de 5.100 anos 14C A.P., foram detectadas pelo aumento nos valores de T.A.C. (10-12 g. m- 2 ano-1), aumento nas taxas de acumulação de massa (T.A.M.) e deposição de sedimentos mais grossos, acompanhados de aumento na tendência continental da natureza da matéria orgânica. As maiores tendências continentais da matéria orgânica sedimentar, registradas aproximadamente entre 1.500-1.200 anos cal. A.P., podem ser decorrentes de uma oscilação negativa do nível do mar. No geral, todos os resultados estão de acordo com as curvas de variação do nível relativo do mar, já descritas na literatura. As análises sedimentares da região estudada revelaram um evento regressivo-trangressivo, que ainda não havia sido referido para o Estado de São Paulo. A partir de ~2.000 anos cal. A.P., o mar teria atingido um nível mais baixo que o atual, com o mínimo provavelmente entre 1.500-1.200 anos cal. A.P., e encontra-se em ascensão até os dias atuais. As características anóxicas e a matéria orgânica sedimentar de origem planctônica registradas junto à praia do Lázaro, em meio a condições de energia relativamente mais alta, em ~900 anos cal. A.P., foram consideradas como resultado de condições óxicas das águas, acompanhadas de altas taxas de produtividade, e subseqüente degradação da matéria orgânica, existentes no período anterior. / Changes in the organic carbon accumulation rates (Corg A.R.) and organic matter source characteristics and their relation to climate and relative sea level changes during the last 8.500 cal. yr. B.P. were investigated in two sediment cores from 2 embayments - Flamengo and Fortaleza inlets - northern coast of São Paulo State. Sediment samples were analysed for grain size, organic Carbon, total Nitrogen and Sulphur contents, and ?13C ratio. Radiocarbon datings provided a model age to the sedimentary deposits. The warm climate interval 8.000-6.000 cal. yr. B.P. was accompanied by an increase in the Corg A.R. (10-13 g.m.-2 yr -1) and a decrease (~2 g.m.-2 yr -1) in the last ~5.000 cal. yr. B.P. Regional climate changes, as stormy weather, that may have ocurred near the maximum transgressive period called Santos Transgression at 5.100 yr. B.P., were recorded in the sedimentary profiles by relatively hight T.A.C. and T.A.M. values, coaser grain size sediment delivery, and enhanced terrestrial organic matter source characteristics. Improvement of terrestrial origin of the sedimentary organic matter recorded between 1.500-1.200 cal. yr. B.P in the studied area was mainly related to a negative sea-level oscillation. All the results are in general agreement with the existing sea level change curves. The sedimentary profiles suggest a negative sea-level fall event, that had\'nt been described yet for São Paulo State. It\'s proposed that the negative oscillation would have begun at ~2.000 cal. yr. B.P.and reached a minimum probably between 1.500-1200 cal. yr. B.P. The relative sea level has been rising since then. Anoxic conditions accompanied by algal origins of sedimentary organic matter, observed in the sedimentary profile from Fortaleza inlet at 900 cal. yr. B.P, during a more oxidizing environment are considered as a result of a previous well oxygencontaining water column and high productivity rates.

No description available.

Essa dissertação pretende esclarecer o estatuto das orações relativas do karitiana (tupi- Arikém). Orações relativas podem ser classificadas como relativas de núcleo externo (RNE) e de núcleo interno (RNI), sendo o principal critério de diferenciação a posição do núcleo em relação à relativa: relativas com o núcleo adjacente à oração subordinada são classificadas como RNEs, enquanto aquelas com o núcleo interno à subordinada são RNIs (DE VRIES, 2006, CULY, 1990). Outro critério utilizado é a marcação de caso: se o núcleo estiver marcado com o caso exigido pelo verbo da matriz, a relativa será uma RNE; se for aquele exigido pelo verbo da relativa, ela será uma RNI (COLE, 1987). Dentro desse quadro, as orações do karitiana são difíceis de classificar: por um lado, o núcleo aparece sempre deslocado para a periferia esquerda (STORTO, 1999), algo característico de RNEs; por outro, a marcação de caso no núcleo segue o padrão de RNIs. À luz do trabalho de Basilico (1996), hipotetizamos que as relativas do karitiana seriam RNIs com frontalização opcional do núcleo. Se for o caso, é esperado que relativas com núcleos não frontalizados sejam permitidas. Montamos então um experimento para verificar se as relativas poderiam ter seus núcleos não frontalizados e testamos 14 falantes com uma metodologia de produção elicitada. Os resultados mostram que, embora haja uma tendência pela frontalização, núcleos não frontalizados são permitidos na língua, pois há casos de relativas de sujeito com a ordem OSV e de relativas de objeto SOtiV, OSV e SOV. Também foram produzidas relativas de objeto sem o morfema de foco do objeto , indicando que ele não é imprescindível para a relativização. Esse quadro aproxima nossas relativas das RNIs, pois RNEs não podem ter núcleos em outras posições além da periferia da oração relativa. Também analisamos propostas de análise sintática para as diversas ordens de palavras coletadas em nosso experimento. Vemos que aquelas que assumem algum tipo de deslocamento do núcleo para Spec de uma projeção de periferia seja CP ou AspP incorrem em diversos problemas, como a impossibilidade de derivar relativas de objeto SOtiV e a incapacidade de excluir estruturas agramaticais com advérbios. Assim, propomos que a frontalização do núcleo é uma adjunção a AspP. Dados de orações relativas com advérbios nos levam ainda a postular que, em relativas de objeto, a frontalização do núcleo ocorre em duas etapas: primeiro um movimento para Spec, vP e, em seguida, a frontalização para uma posição de adjunto de AspP. Essa primeira etapa do movimento seria marcada pela presença de em v e estaria na base do sincretismo desse morfema, que também está presente em perguntas qu- de objeto e em construções de foco do objeto. Por fim, oferecemos ainda uma análise da correlação entre a presença do morfema e a frontalização do objeto a partir do modelo de fases de Chomsky (2000, 2001), admitindo que o movimento do objeto para a borda do sintagma verbal seria uma operação sintática imprescindível para a subsequente frontalização do objeto / This dissertation aims to clarify the status of relative clauses in Karitiana (Tupi-Arikém). Relative clauses are traditionally classified as externally (EHRC) or internally-headed (IHRC) and the main criterion for their differentiation is the heads position: EHRCs have their heads adjacent to the relative clause itself, whereas IHRCs have internal heads (DE VRIES, 2006, CULY, 1990). Another criterion is case-marking: if the head is marked with the case demanded by the matrix verb, the relative is an EHRC; if it is the one demanded by the embedded verb, it will be an IHRC (COLE, 1987). Within this framework, karitiana relative clauses are hard to classify: on one hand, the head always appears fronted to the left periphery (STORTO, 1999), which resembles the pattern found in EHRCS; on the other hand, the case-marking on the head is similar to IHRCs. In the light of Basilicos (1996) work, one can hypothesize that karitiana relative clauses are IHRCs with optional head frontalization. In this case, it is expected that relatives with non-fronted heads will be allowed in the language. An experiment was designed in order to verify if karitiana relatives could have non-fronted heads and 14 speakers were tested with an elicited production methodology. The results show that, although there is a preference for frontalization, non-fronted heads are possible in the language, since subject relatives OSV and object relatives SOtiV, OSV and SOV were produced. There are also cases of object relatives without the object focus morpheme , indicating that it is not indispensable in relative clause formation. These results bring karitiana relatives closer to IHRCs, because EHRCs cannot have their heads in any other positions than in the periphery of the clause. We also discuss some syntactic proposals for the word orders found in our experiment, claiming that those which assume head dislocation to Spec of CP and AspP face some problems, such as the derivation of SOtiV object relatives and ungrammatical structures with adverbs. Therefore, our proposal is that the frontalization of the head is an adjunction to AspP. Paradigms of relative clauses with adverbs also show that, in object relatives, the frontalization of the head occurs in two steps: first the head moves to Spec, vP and then it is further fronted to the position of adjunct of Spec, AspP. The first step is marked with on v and it underlies the syncretism of this morpheme, which is also present in object wh- questions and object focus constructions. Finally, the correlation between and the frontalization of the head is analyzed within the phase theory framework (CHOMSKY, 2000, 2001) and it is assumed that object movement to vPs edge is a syntactic requirement for further frontalization