abstract: It is common in the analysis of data to provide a goodness-of-fit test to assess the performance of a model. In the analysis of contingency tables, goodness-of-fit statistics are frequently employed when modeling social science, educational or psychological data where the interest is often directed at investigating the association among multi-categorical variables. Pearson's chi-squared statistic is well-known in goodness-of-fit testing, but it is sometimes considered to produce an omnibus test as it gives little guidance to the source of poor fit once the null hypothesis is rejected. However, its components can provide powerful directional tests. In this dissertation, orthogonal components are used to develop goodness-of-fit tests for models fit to the counts obtained from the cross-classification of multi-category dependent variables. Ordinal categories are assumed. Orthogonal components defined on marginals are obtained when analyzing multi-dimensional contingency tables through the use of the QR decomposition. A subset of these orthogonal components can be used to construct limited-information tests that allow one to identify the source of lack-of-fit and provide an increase in power compared to Pearson's test. These tests can address the adverse effects presented when data are sparse. The tests rely on the set of first- and second-order marginals jointly, the set of second-order marginals only, and the random forest method, a popular algorithm for modeling large complex data sets. The performance of these tests is compared to the likelihood ratio test as well as to tests based on orthogonal polynomial components. The derived goodness-of-fit tests are evaluated with studies for detecting two- and three-way associations that are not accounted for by a categorical variable factor model with a single latent variable. In addition the tests are used to investigate the case when the model misspecification involves parameter constraints for large and sparse contingency tables. The methodology proposed here is applied to data from the 38th round of the State Survey conducted by the Institute for Public Policy and Michigan State University Social Research (2005) . The results illustrate the use of the proposed techniques in the context of a sparse data set. / Dissertation/Thesis / Ph.D. Mathematics 2011

La biologie structurale est la branche de la biologie qui étudie la structure et l'organisation spatiale des macromolécules.La biologie structurale concerne en particulier la détermination à l'échelle atomiquede la structure 3D, aux changement de conformation des macromolécules, et à la dynamique de ces structures.De nos jours, les techniques expérimentales modernes telles que la résonance magnétique nucléaire, la cristallographie aux rayons X et plus récemmentla microscopie cryoélectronique peuvent produire des cartes de densité à haute résolution, qui combinées aux informations sur la séquence d'une moléculepermettent aux biologistes de résoudre les structures 3D de la molécule à l'étude.Cependant, dans certains cas, la résolution des cartes de densité n'est pas suffisante.Dans un tel cas, on alignegénéralement des sous-unités individuelles, obtenues à haute résolution, dans la carte de densité de base résolution.Mentionnons qu'il est également également possible de déterminer la structure 3D d'un assemblage biologique en ancrant plusieurs sous-unités ensemble.C'est cependant un problème beaucoup plus difficile.Ces problèmes d'amarrage et d'alignement peuvent être formulés comme un problème d'optimisation dont la fonction de coût est écrite comme la corrélation croisée de deux autres fonctions.Les algorithmes d'ancrage originaux ont été formulés comme des problèmes de "clé et verrou", dans lesquels les protéines étaient considérées comme des corps rigides.Il est cependant naïf de considérer les macromolécules comme des corps rigides. Les protéines sont flexibles et peuventsubir de grands changements conformationnels lors de la liaison à d'autres molécules. Considérer les problèmes d'ancrage comme des problèmes de"clé et verrou" n'est donc pas suffisant.Une méthode d'ancrage flexible standard utilise donc l'approche "aligner puis affiner", qui, dans certains cas, peut omettre de bonnes conformations.Cette thèse se concentre sur deux axes principaux.Le premier axe est le développement d'une nouvelle méthode qui échantillonne de manière exhaustive les mouvements de corps rigides et les mouvements collectifs, calculés par analyse en modes propres (AMP).Nous présentons d'abord une méthode qui utilise la transformée de Fourier rapide pour échantillonner une approximation quadratique de la fonction coût. Ensuite, la méthode effectuela recherche flexible en maximisant l'approximation quadratique de la fonction de coût dans un certain domaine de recherche. Cette méthode garantit de trouver la meilleure conformation flexible.Nous présentons ensuite une version en itterative de notre algorithme, qui trouve les mouvements collectifs qui maximisent le score d'amarrage par rapport aux degrés de liberté (DDLs) du corps rigide.La méthode échantillonne de manière exhaustive à la fois les mouvements de corps rigides et les mouvements collectifs en maximisant le maximum lisse selon les DDls correspondant aux transformations rigides de la fonction coût.Les deux méthodes ont été appliquées à des problèmes d'alignement sur des exemples réels et artificiels.De plus, nous présentons un exemple dans lequel l'approche "aligner puis raffiner" n'est pas capable de trouver la bonne conformation tandis quenotre méthode peut trouver ladite conformation.Le deuxième axe est le développement d'une nouvelle extrapolation des mouvements calculés par l'AMP.Nous montrons qu'il est possible, avec des calculs minimaux, d'extrapoler les mouvements instantanés calculés par l'AMP dans le sous espaces des rotations-traslations des blocs (RTB) comme une rotationpresque pure autour d'un certain axe.Nous avons appliqué cette méthode appelée NOLB sur différents systèmes biologiques et avons pu, d'une part, récupérer des mouvements biologiquement pertinents et d'autre part démontrer que la méthode NOLB génère des structures avec une meilleure topologie qu'une méthode d'AMP linéaire. / Structural biology is a branch of molecular biology, biochemistry, and biophysics concerned with the molecular structure of macromolecules, how they acquire the structures they have,and how alterations in their structures affect their function.These molecules are a topic of interest because they serve to keep the cellsalive and functioning.Nowadays, modern experimental techniques, such as nuclear magnetic resonance (NMR), X-ray crystallography and more recently cryo-electron microscopy (cryo-EM) canproduce high resolution density maps, which combined with the information about the sequence of a molecule allows biologists to solve thethree-dimensional (3D) structures of the molecule under study. However, when studding large biological assemblies, experimental techniques are notalways able to generate density maps with a high enough resolution. In such a case, one typically fits individual sub-units, which weresolved using at a higher resolution, into the lower-resolution density map.Let us also mention that it is also possible determine the 3D structure of a biological assembly by docking several sub-units together.This is a much more difficult problem though.These docking and fitting problems can be reformulated as an optimization problem whose cost function can be written as the cross-correlation of two functions.The first fitting and docking algorithms were formulated as "lock and key" problems, in which the proteins were considered as rigid body.However, considering macromolecules, especially proteins, as rigid bodies is not realistic.Proteins are indeed flexible and can undergo large conformational changesupon binding to other molecules.Considering docking and fitting problems as "lock and key" problems is therefore not sufficient.Therefore, a standard flexible docking/fitting method first uses a six-dimensional (6D) rigid body docking/fitting algorithm and then flexibly relaxes the top docking/fitting poses.This approach will be thus refereed to as to the fit then refine approach.However, in some cases, such an approach can miss good conformations.This thesis focuses on two main axes.The first axis is the development of a new method that exhaustively samples both rigid-body and collective motions computed via normal mode analysis (NMA).We first present a method that combines the advantages of the Fourier transform (FFT)-based exhaustive search, which samples all the conformations of a system under study on a grid, with a local optimization technique thatguarantees to find the nearest optimal off-grid and flexible conformation.The algorithm first samples a quadratic approximation of a scoring function on a 6D grid. Then, the method performs the flexible search by maximizing the quadratic approximation of the cost functionwithin a certain search space.We then present a multi-step version of our algorithm, which finds the collective motions that maximize the docking score with respect to the rigid-body degrees of freedom (DOFs).The method exhaustively samples both rigid-body and collective motions by maximizing the soft maximum over the rigid body DOFs of the docking/fitting cost function.Both methods were applied to docking problems on both real and artificial example and we were able to design a benchmark in which the fit then refine approach fails at finding the correct conformation whileour method succeeds.The second axis is the development of a new extrapolation of motions computed by NMA.We show that it is possible, with minimal computations, to extrapolate the instantaneous motions computed by NMA in the the rotations-translations of blocks (RTB) subspace as an almost pure rotation around a certain axis.We applied this non-linear block (NOLB) method on various biological systems and were able to, firstly, retrieve biologically relevant motions andsecondly, to demonstrate that the NOLB method generates structures with a better topology than a linear NMA method.

L'assemblage de pièces de structures composites, notamment en carbone/époxy, est souvent réalisé par liaison boulonnée. Ceci nécessite le perçage des logements de fixations. Cette opération de perçage doit répondre à des contraintes de fiabilité, de productivité et de qualité d'usinage. L'opération de perçage doit donc être maitrisée, ce qui passe par une meilleure compréhension des phénomènes présents en perçage. Mais l'étude de ces phénomènes se heurte à des obstacles tels que l'aspect confiné de l'opération, la géométrie complexe et variable du foret, la variation de vitesse le long de l'arête principale de coupe... Afin de s'affranchir de ces obstacles, des essais élémentaires représentatifs de la coupe en perçage peuvent être proposés. Ce travail de thèse s'inscrit dans cette optique. Ce travail s'articule autour de trois axes. Le premier concerne une étude de la représentativité des différents essais élémentaires vis-à-vis du perçage. Un outil d'identification de géométrie d'outil est proposé ; il permet d'identifier l'évolution de la géométrie locale des outils coupants et ainsi de proposer des essais élémentaires qui soient géométriquement et cinématiquement représentatifs de la coupe en perçage. Le second axe développé propose une étude des phénomènes présents en coupe orthogonale quasi-statique. Cela permet de s'affranchir des effets liés à la vitesse. Dans ce cadre, des essais de corrélation d'images en coupe orthogonale sont proposés. Ils permettent l'analyse des champs de déplacements et de déformations. Une analyse des efforts générés, de la morphologie des copeaux et des états de surfaces obtenus en fonction de l'angle de coupe et de l'angle ?2 entre la vitesse de coupe et la direction des fibres est également réalisée. Deux types de modèles numériques, macro- mécanique et micro-mécanique, sont proposés. Ils sont confrontés aux résultats expérimentaux. Le dernier axe de travail présente l'analyse des phénomènes liés à la vitesse de coupe en coupe orthogonale. Le modèle macro-mécanique est modifié afin d'intégrer des phénomènes tels que la variation des contraintes à rupture de la matrice en fonction de la vitesse de déformation, ou encore l'évolution du frottement en fonction de la vitesse de glissement. Les résultats du modèle sont confrontés aux résultats expérimentaux. / In aeronautical sector, assembly of CFRP composite structures requires the drilling of the fastener holes. Requirements of reliability, productivity and machining quality are imposed on the drilling process. The operation must be mastered, which requires a better understanding of the phenomena occurring during drilling. But the study of these phenomena faces several major challenges such as the confined aspect of the operation, the complex and variable geometry of the cutting tool, the speed variation along the main cutting edge... To overcome these obstacles, elementary testing representative of the drilling cutting phenomena could be implemented. This is the purpose of this work, which is based on three axes. The first is a study of the geometrical and kinematic representativeness of the elementary tests regarding the drilling operation. A tool geometry identification program is developed. It identifies the evolution of the local geometry of cutting tools along the main cutting edges and allows defining the elementary tests that are geometrically and kinematically representative of the drilling cutting. The second axis developed proposes a study of the phenomena occuring in quasi-static orthogonal cutting. Thus, the effects relative to the cutting speed are not considered. In this context, digital image correlation tests during orthogonal cutting are conducted. They lead to the analysis of the displacements and strains fields. Generated forces, chip morphology and machined surface texture are also investigated, in relation to the rake angle and the angle ?2 between the cutting speed and the direction of fibres. For this configuration, two types of numerical models, a macro-mechanical and a micro-mechanical one, are developed. A comparison is made in relation to experimental results. The last axis of this work concern the analysis of phenomena related to the cutting speed during orthogonal cutting. The macro-mechanical model is modified to include such phenomena as the variation of the matrix breaking stress as a function of strain rate, or the evolution of the friction coefficient according to the sliding velocity. The model results are compared to experimental results.

Made available in DSpace on 2014-06-11T19:26:07Z (GMT). No. of bitstreams: 0 Previous issue date: 2008-07-28Bitstream added on 2014-06-13T20:14:40Z : No. of bitstreams: 1 sesana_rc_me_jabo.pdf: 242011 bytes, checksum: 2b8f05f5410e0ff737c41f94ef98b886 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Foram estimados parâmetros genéticos da produção de leite acumulada até os 305 dias de lactação (P305) de 1.946 búfalas da raça Murrah no decorrer da idade ao parto utilizando os modelos de repetibilidade, multi-característica e de regressão aleatória (MRA), e correlações de ordem entre os valores genéticos para a P305 nas diferentes idades ao parto preditos pelos modelos de regressão aleatória e de repetibilidade e para a P305 acumulada até os 8 anos de idade, considerando diferentes intensidades seletivas, número de filhas e sua distribuição nos rebanhos. Também foram estimados parâmetros genéticos para a produção de leite no dia do controle (PLDC) de 1.433 primeiras lactações de búfalas da raça Murrah utilizando MRA. Os modelos de repetibilidade e multi-característica incluíram para a P305 os efeitos fixos de grupo de contemporâneos composto por rebanho, ano e estação do parto, número de ordenhas (1 ou 2 ordenhas diárias) e o efeito linear e quadrático da covariável idade da vaca ao parto e os efeitos aleatórios de animal, ambiente permanente e residual, com exceção do efeito de ambiente permanente para o segundo modelo. No MRA, as análises tanto para a P305 quanto para a PLDC foram realizadas por meio de um modelo uni-característica de regressão aleatória, considerando os mesmos efeitos aleatórios e fixos dos modelos de repetibilidade e multi-característica. No entanto, para a PLDC o GC foi composto por rebanho, ano e mês do controle. Uma regressão ortogonal de terceira ordem foi usada para modelar a trajetória média da população e os efeitos genéticos aditivos e de ambiente permanente. O MRA considerando um polinômio de Legendre de terceira ordem para os efeitos genético aditivo e de ambiente permanente e uma função de variância de segunda ordem (3.3.fv2) foi o mais adequado para o ajuste... / Genetic parameters were estimated for accumulated 305-day milk yields (P305) of 1,946 Murrah buffaloes in different ages of calving using repeatability, multi-trait and random regression models (RRM) and Spearman correlation among the breeding values for P305 in different ages of calving predicted using RRM and repeatability models and for accumulated P305 until 8 years old, considering different selective intensities, number of daughters and its distribution in herds. Were also estimated genetic parameters for first lactation test day milk yields (PLDC) of 1,433 Murrah buffaloes using RRM. repeatability and multi-trait models for the P305 included the fixed effects of contemporary group, composed by herd, year and season of calving, milking frequency (1 or 2), age at calving as covariable with linear and quadratic effect and animal, permanent environmental and residual random effects, with exception of the permanent environmental effect for the second model. In the RRM, the analyses for the P305 and PLDC were both achieved through a uni-trait model of random regression, included the same random effects and fixed effects of the repeatability and multi-trait models. However, for the PLDC the contemporary group was composed by herd, year and month of test. A third order regression on Legendre orthogonal polynomial of milk yields was used to model the population mean trend and the additive genetic and permanent environmental regressions. The RRM with a third order covariance function for genetic and permanent environmental effects and a second order variance function (3.3.fv2) was indicated as the best for P305. Heritability estimated for RRM to P305 ranges from 0.19 (6 years) to 0.23 (11 years) and for multi-trait ranged from 0.13 (6 years) to 0.36 (4 years). Heritability estimated for repeatability model was 0.20. Genetics and phenotypic correlation among the P305 in different ages... (Complete abstract click electronic access below)

One of the primary tasks of the mobile system designers in order to support high density of devices in a CDMA system is to create a code sequence with a capacity for large number of spreading code sequences having low cross-correlation values between them, in order to ensure accommodation of large number of users and to minimise the effect of multiple access interference. In this research, the design for a novel hybrid orthogonal very large set (HOVLS) code sequence is proposed for high density mobile application scenarios. The design and development of both fixed and variable spreading factor code sequences are presented in this thesis. Both type of code sequences have been implemented via simulation using Matlab/Simulink. The performance of the code sequences has been evaluated and compared with that of existing code sequences. The proposed code sequences are more advantageous for high density mobile networks. The unique feature of the fixed length HOVLS code sequence is that its ACF, CCF, and BER performances are similar to that of orthogonal Gold code sequence and orthogonal m-sequence under Rayleigh flat and frequency selective fading channel conditions while having a significantly higher capacity than those orthogonal code sequences. The proposed HOVLS code sequence could support 134 different cells which is more than twice than that of orthogonal Gold code sequence and orthogonal m-sequence. To the knowledge of the author, this is the largest reported family size in the literature for an orthogonal code sequence for CDMA applications. In order to support variable data rate, fixed length HOVLS code sequence was developed into orthogonal variable spreading factor code sequence. It is shown that the proposed OVSF code sequence has slightly better CCF than those of OVSF Gold code sequence and m-sequence in terms of CM (correlation margin). The ACF of the proposed OVSF code sequence is similar to those of OVSF Gold code sequence and m-sequence. The proposed OVSF code sequence possesses comparable BER performance to those of orthogonal Gold code sequence and orthogonal m-sequence under flat fading channel condition. Whereas, the BER performance of the proposed OVSF code sequence is slightly better than that of Gold code sequence and OVSF m-sequence under frequency selective fading channel. Therefore, the proposed HOVLS code sequence is appropriate code sequence in CDMA systems than those of orthogonal Gold code sequence and orthogonal m-sequence for both fixed and variable rate high density network applications.

Made available in DSpace on 2014-06-11T19:27:14Z (GMT). No. of bitstreams: 0 Previous issue date: 2004-07-29Bitstream added on 2014-06-13T19:55:34Z : No. of bitstreams: 1 santos_kac_me_ilha.pdf: 2835945 bytes, checksum: c558fa3df558636f50b76fdd657c5ff5 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Nas técnicas de identificação de parâmetros, procuram-se determinar os valores desconhecidos pela manipulação dos sinais de entrada e saída do sistema. O tratamento e análise de sinais são relativamente recentes na engenharia, sendo que seu desenvolvimento deu-se juntamente com o dos sensores e condicionadores de sinais e mais recentemente, com os sistemas automáticos de aquisição de dados. Vários métodos têm sido propostos para resolver problemas de identificação, embora nenhum deles possa ser considerado como sendo universalmente adequado a todas as situações. Conhecendo-se os parâmetros dos sistemas, pode-se acompanhar através de monitoramento e técnicas de identificação, a evolução de possíveis falhas devido à variação destes parâmetros. Os processos de identificação, a partir de funções ortogonais, começam com a construção de uma matriz operacional, o que permite, através de integrações a conversão de um conjunto de equações diferenciais em um conjunto de equações algébricas e consequentemente a obtenção dos parâmetros desconhecidos. Neste trabalho, apresentam-se as técnicas de Identificação de Parâmetros utilizando as funções ortogonais de Fourier e polinomiais de Legendre e Chebyshev. / In the parameter identification techniques, it is important to determine the unknown values in the manipulation of input and output signal of the system. The treatment and analysis of signals are relatively recent in the engineering, and its development took place with the sensors and the signal conditioning and recently, with the automatic data acquisition systems. Various methods have been proposed to solve identification problems, although any of them can be regarded universally adequate to all the situations. If the parameters of the systems, is known it can be accompanied, through monitoring and identification techniques, the evolution of possible fault due to the variation of the parameters. The identification process, from these types of functions, start with the construction of an operational matrix for the integration of orthogonal bases vectors, which allow the conversion of a differential equation set to a algebraic equation set, obtaining the unknown parameters. In this work, the parameter identification techniques used, the orthogonal functions of Fourier and polynomial of Legendre and Chebyshev, is presented.

Orientador: Dimitar Kolev Dimitrov / Banca: Carlos Tomei / Banca: José  Alberto Cuminato / Resumo: Neste trabalho estudaremos a relação existente entre polinômios ortogonais e matrizes aleatórias. Exibiremos uma caracterização de polinômios ortogonais via problemas de Riemann-Hilbert, a qual tem se mostrado uma ferramenta única para obtenção de assintóticas de polinômios ortogonais. Posteriormente, estudaremos a teoria básica dos ensembles unitários de matrizes aleatórias. Por fim, mostraremos como a teoria de assintóticas de polinômios ortogonais pode ser usada na análise assintótica de estatísticas de matrizes aleatórias, nos levando a resultados de universalidade para os ensembles unitários / Abstract: We will exhibit a characterization of orthogonal p olynomials via Riemann-Hilbert problems, which has been shown a powerful to ol for studying asymptotics of orthogonal polynomials. Posteriorly we will review the basic theory of unitary ensembles of random matrices. At the end, we will show how asymptotics of orthogonal polynomials can be used to study asymptotics of several statistics in random matrix theory, obtaining universality results for the unitary ensembles / Mestre

Resumo: Foram estimados parâmetros genéticos da produção de leite acumulada até os 305 dias de lactação (P305) de 1.946 búfalas da raça Murrah no decorrer da idade ao parto utilizando os modelos de repetibilidade, multi-característica e de regressão aleatória (MRA), e correlações de ordem entre os valores genéticos para a P305 nas diferentes idades ao parto preditos pelos modelos de regressão aleatória e de repetibilidade e para a P305 acumulada até os 8 anos de idade, considerando diferentes intensidades seletivas, número de filhas e sua distribuição nos rebanhos. Também foram estimados parâmetros genéticos para a produção de leite no dia do controle (PLDC) de 1.433 primeiras lactações de búfalas da raça Murrah utilizando MRA. Os modelos de repetibilidade e multi-característica incluíram para a P305 os efeitos fixos de grupo de contemporâneos composto por rebanho, ano e estação do parto, número de ordenhas (1 ou 2 ordenhas diárias) e o efeito linear e quadrático da covariável idade da vaca ao parto e os efeitos aleatórios de animal, ambiente permanente e residual, com exceção do efeito de ambiente permanente para o segundo modelo. No MRA, as análises tanto para a P305 quanto para a PLDC foram realizadas por meio de um modelo uni-característica de regressão aleatória, considerando os mesmos efeitos aleatórios e fixos dos modelos de repetibilidade e multi-característica. No entanto, para a PLDC o GC foi composto por rebanho, ano e mês do controle. Uma regressão ortogonal de terceira ordem foi usada para modelar a trajetória média da população e os efeitos genéticos aditivos e de ambiente permanente. O MRA considerando um polinômio de Legendre de terceira ordem para os efeitos genético aditivo e de ambiente permanente e uma função de variância de segunda ordem (3.3.fv2) foi o mais adequado para o ajuste... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Genetic parameters were estimated for accumulated 305-day milk yields (P305) of 1,946 Murrah buffaloes in different ages of calving using repeatability, multi-trait and random regression models (RRM) and Spearman correlation among the breeding values for P305 in different ages of calving predicted using RRM and repeatability models and for accumulated P305 until 8 years old, considering different selective intensities, number of daughters and its distribution in herds. Were also estimated genetic parameters for first lactation test day milk yields (PLDC) of 1,433 Murrah buffaloes using RRM. repeatability and multi-trait models for the P305 included the fixed effects of contemporary group, composed by herd, year and season of calving, milking frequency (1 or 2), age at calving as covariable with linear and quadratic effect and animal, permanent environmental and residual random effects, with exception of the permanent environmental effect for the second model. In the RRM, the analyses for the P305 and PLDC were both achieved through a uni-trait model of random regression, included the same random effects and fixed effects of the repeatability and multi-trait models. However, for the PLDC the contemporary group was composed by herd, year and month of test. A third order regression on Legendre orthogonal polynomial of milk yields was used to model the population mean trend and the additive genetic and permanent environmental regressions. The RRM with a third order covariance function for genetic and permanent environmental effects and a second order variance function (3.3.fv2) was indicated as the best for P305. Heritability estimated for RRM to P305 ranges from 0.19 (6 years) to 0.23 (11 years) and for multi-trait ranged from 0.13 (6 years) to 0.36 (4 years). Heritability estimated for repeatability model was 0.20. Genetics and phenotypic correlation among the P305 in different ages... (Complete abstract click electronic access below) / Orientador: Humberto Tonhati / Coorientadora: Lúcia Galvão de Albuquerque / Coorientadora: Lenira El Faro Zadra / Banca: Danísio Prado Munari / Banca: Maria Eugênia Zerlotti Mercadante / Mestre

Orientador: Gilberto Pechoto de Melo / Banca: Cleudmar Amaral de Araújo / Banca: Vicente Lopes Júnior / Resumo: Nas técnicas de identificação de parâmetros, procuram-se determinar os valores desconhecidos pela manipulação dos sinais de entrada e saída do sistema. O tratamento e análise de sinais são relativamente recentes na engenharia, sendo que seu desenvolvimento deu-se juntamente com o dos sensores e condicionadores de sinais e mais recentemente, com os sistemas automáticos de aquisição de dados. Vários métodos têm sido propostos para resolver problemas de identificação, embora nenhum deles possa ser considerado como sendo universalmente adequado a todas as situações. Conhecendo-se os parâmetros dos sistemas, pode-se acompanhar através de monitoramento e técnicas de identificação, a evolução de possíveis falhas devido à variação destes parâmetros. Os processos de identificação, a partir de funções ortogonais, começam com a construção de uma matriz operacional, o que permite, através de integrações a conversão de um conjunto de equações diferenciais em um conjunto de equações algébricas e consequentemente a obtenção dos parâmetros desconhecidos. Neste trabalho, apresentam-se as técnicas de Identificação de Parâmetros utilizando as funções ortogonais de Fourier e polinomiais de Legendre e Chebyshev. / Abstract: In the parameter identification techniques, it is important to determine the unknown values in the manipulation of input and output signal of the system. The treatment and analysis of signals are relatively recent in the engineering, and its development took place with the sensors and the signal conditioning and recently, with the automatic data acquisition systems. Various methods have been proposed to solve identification problems, although any of them can be regarded universally adequate to all the situations. If the parameters of the systems, is known it can be accompanied, through monitoring and identification techniques, the evolution of possible fault due to the variation of the parameters. The identification process, from these types of functions, start with the construction of an operational matrix for the integration of orthogonal bases vectors, which allow the conversion of a differential equation set to a algebraic equation set, obtaining the unknown parameters. In this work, the parameter identification techniques used, the orthogonal functions of Fourier and polynomial of Legendre and Chebyshev, is presented. / Mestre

Orientador: Alagacone Sri Ranga / Banca: Sandra Augusta Santos / Banca: Cleonice Fátima Bracciali / Resumo: Dentre as diversas fórmulas de quadratura interpolatórias estão aquelas que utilizam em sua construção as propriedades dos polinômios ortogonais Pn, ou ainda dos polinômios similares Bn. Consideramos, aqui, fþormulas de quadratura envolvendo polinôomios em x da forma .n(x, .) = Pn-1(.)Pn(x) - Pn(.)Pn-1(x), e da forma Gn(x, u) = Bn-1(u)Bn(x) - Bn(u)Bn-1(x). Abordamos ainda certas fþormulas de quadratura que visam aproximar a integral de um produto de duas funções k e f sendo k Lebesgue integrþavel e f Riemann integrþavel. O principal objetivo deste trabalho þe analisar propriedades das fþormulas de quadratura utilizando-se .n e obter propriedades anþalogas para o caso onde utiliza-se Gn, bem como estudar o erro e as propriedades de convergência das fórmulas envolvendo k e f. Propriedades dos pesos das fórmulas de quadratura nos diversos casos são analisadas, a convergência das fórmulas associadas a k e f são estudadas mediante determinadas escolhas de pontos. / Abstract: Among the many well known quadrature formulas one finds those interesting interpolatory quadrature formulas that take advantage of the properties of orthogonal polynomials Pn or similar polynomials Bn. Here, we consider the interpolatory quadrature rules based on the zeros of the polynomials øn(x, î) = Pn.1(î)Pn(x).Pn(î)Pn.1(x), and Gn(x, u) = Bn.1(u)Bn(x) . Bn(u)Bn.1(x) where î and u are arbitrary parameters. One of the objective of this dissertation is to study some of the known properties of quadrature rules based on øn(x, î) and consider the analogous properties of the quadrature rules based on Gn(x, u).We also look at the convergence properties of those quadrature rules that serve to approximate integrals of the product of functions k and f, where k is a Lebesgue integrable function and f needs to be a Riemann integrable function. / Mestre