This thesis is about high–dimensional problems considered under the so{called Kolmogorov condition. Hence, we consider research questions related to random matrices with p rows (corresponding to the parameters) and n columns (corresponding to the sample size), where p &gt; n, assuming that the ratio <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Csmall%5Cfrac%7Bp%7D%7Bn%7D" /> converges when the number of parameters and the sample size increase. We focus on the eigenvalue distribution of the considered matrices, since it is a well–known information–carrying object. The spectral distribution with compact support is fully characterized by its moments, i.e., by the normalized expectation of the trace of powers of the matrices. Moreover, such an expectation can be seen as a free moment in the non–commutative space of random matrices of size p x p equipped with the functional <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Csmall%20%5Cfrac%7B1%7D%7Bp%7DE%5BTr%5C%7B%5Ccdot%5C%7D%5D" />. Here, the connections with free probability theory arise. In the relation to that eld we investigate the closed form of the asymptotic spectral distribution for the sum of the quadratic forms. Moreover, we put a free cumulant–moment relation formula that is based on the summation over partitions of the number. This formula is an alternative to the free cumulant{moment relation given through non{crossing partitions ofthe set. Furthermore, we investigate the normalized <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Csmall%20E%5B%5Cprod_%7Bi=1%7D%5Ek%20Tr%5C%7BW%5E%7Bm_i%7D%5C%7D%5D" /> and derive, using the dierentiation with respect to some symmetric matrix, a recursive formula for that expectation. That allows us to re–establish moments of the Marcenko–Pastur distribution, and hence the recursive relation for the Catalan numbers. In this thesis we also prove that the <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Csmall%20%5Cprod_%7Bi=1%7D%5Ek%20Tr%5C%7BW%5E%7Bm_i%7D%5C%7D" />, where <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Csmall%20W%5Csim%5Cmathcal%7BW%7D_p(I_p,n)" />, is a consistent estimator of the <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Csmall%20E%5B%5Cprod_%7Bi=1%7D%5Ek%20Tr%5C%7BW%5E%7Bm_i%7D%5C%7D%5D" />. We consider <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Csmall%20Y_t=%5Csqrt%7Bnp%7D%5Cbig(%5Cfrac%7B1%7D%7Bp%7DTr%5Cbig%5C%7B%5Cbig(%5Cfrac%7B1%7D%7Bn%7DW%5Cbig)%5Et%5Cbig%5C%7D-m%5E%7B(t)%7D_1%20(n,p)%5Cbig)," />, where <img src="http://www.diva-portal.org/cgi-bin/mimetex.cgi?%5Csmall%20m%5E%7B(t)%7D_1%20(n,p)=E%5Cbig%5B%5Cfrac%7B1%7D%7Bp%7DTr%5Cbig%5C%7B%5Cbig(%5Cfrac%7B1%7D%7Bn%7DW%5Cbig)%5Et%5Cbig%5C%7D%5Cbig%5D" />, which is proven to be normally distributed. Moreover, we propose, based on these random variables, a test for the identity of the covariance matrix using a goodness{of{t approach. The test performs very well regarding the power of the test compared to some presented alternatives for both the high–dimensional data (p &gt; n) and the multivariate data (p ≤ n).

This thesis studies the use of the Source Reconstruction Method (SRM) to characterize antennas. The SRM calculates equivalent sources/currents on an arbitrarily shaped reconstruction surface to represent the original antenna. This is done by enforcing that the original antenna and equivalent currents radiate the same field at user selected measurement locations. These equivalent currents spatially characterize the original antenna because they can be used in direct radiation problems to obtain field estimates anywhere outside the reconstruction surface, including the far-field. First a spherical SRM algorithm is implemented and the diagnostic capabilities of the SRM are also synthetically shown through an example with an array of elementary dipoles. It is then shown that the SRM compares well to pre-existing commercial antenna software over different frequencies and can also be used successfully with a partial dataset. It is demonstrated that the equivalent currents can also provide meaningful information with experimental data. Next the hierarchical matrix framework is studied in conjunction with the SRM to decrease the algorithm's memory requirement and increase the speed of execution. It is shown that it is beneficial to use the hierarchical matrix framework with the SRM when using Love's condition or with measured data on a surface very close to the reconstruction surface. The SRM is then used to obtain incident field estimates in microwave imaging systems. Using a 2D transverse magnetic framework, we show that even with the limited data available in typical microwave tomography setups the SRM can produce incident field estimates in the imaging domain. These estimates are then used along with an MR-GNI algorithm to image synthetic and experimental objects with uncalibrated measured data. / October 2016

The aim of this thesis is to conduct a general investigation in the field of sparse matrices, to investigate and compare various techniques for handling sparse systems suggested in the literature, to develop some new techniques, and to discuss the feasibility of using sparsity techniques in the solution of overdetermined equations and the eigenvalue problem.

The capacity of Raven's Standard Progressive Matrices (SPM) and the Booklet Category Test (BCT) to discriminate between groups of brain-injured, simulated malingering, and normal participants was investigated in this study. Exploratory analyses were also conducted to examine the differences between groups categorized as sophisticated and naive fakers. Clinical decision rules and discriminant function analyses were utilized to identify malingerers. Clinical decision rules ranged in hit rates from 41% to 78%, in sensitivity from 2% to 100%, and in specificity from 86% to 100%. Discriminant functions ranged in hit rates from 81% to 86%, in sensitivity from 68% to 73% and in specificity from 82% to 87%. Overall, the least helpful detection method examined was below chance responding on either measure, while the most efficient was gross errors for SPM.

Thèse numérisée par la Division de la gestion de documents et des archives de l'Université de Montréal.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Depuis une quinzaine d'années de nombreux travaux s'attachent à utiliser les systèmes MIMO afin d'augmenter la capacité de Shannon associée aux traditionnels systèmes SISO. Dans ce but, un problème crucial consiste en la conception de l'émetteur optimal au sens de la capacité de Shannon. Cette problématique a fait l'objet de nombreuses études dans le cas où le canal de transmission MIMO est non sélectif en fréquence ; elle est cependant nettement moins mature dans le cadre d'un canal MIMO sélectif en fréquence. Cette thèse s'intéresse ainsi dans une première partie à l'optimisation, au sens de la capacité ergodique, de la covariance du vecteur transmis, via la théorie des matrices aléatoires. L'utilisation de plusieurs antennes d'émission permet également d'augmenter les performances en réception grâce à la diversité induite. Dans une seconde partie, nous nous intéressons ainsi à la diversité liée aux récepteurs MMSE. A l'inverse des récepteurs ML ces récepteur sont sous-optimaux mais très simples à mettre en oeuvre. Dans un premier temps nous étudions la diversité de tels récepteurs à haut SNR pour des canaux sélectifs en fréquence, tandis que nous nous attardons dans un second temps sur un facteur de diversité, l'utilisation des codes spatio-temporels en bloc, plus spécifiquement l'utilisation du code d'Alamouti. Ainsi, nous proposons et analysons en contexte multi-utilisateur un nouveau récepteur MMSE robuste aux interférences car exploitant au mieux les degrés de liberté du canal / For fifteen years many studies have used MIMO systems to increase the Shannon capacity of the traditional SISO systems. To this end, a crucial problem is the design of transmitters which are optimal w.r.t. Shannon capacity, by the use of space-time codes or of prior knowledge on the transmission channel. These problems have been addressed by many studies in the case of frequency flat MIMO channels but are really less mature for frequency selective MIMO channels. This thesis focuses in the first part on the optimization, w.r.t. the ergodic capacity, of the covariance of the vector transmitted, via the Random Matrix Theory. Using multiple transmit antennas also gives rise to diversity, which improves the receiving performance. In the second part, we thus focus on the diversity, in the specific case of a MMSE receiver. Unlike the ML receiver, this receiver is suboptimal but very simple to implement. We first study the diversity at high SNR for frequency selective channels. We then focus on a diversity factor, the use of space-time codes in block (STBC), specifically the use of the Alamouti code. Thus, we propose and analyze in the multiuser context a new MMSE receiver robust to interference thanks to its ability to use optimally the degrees of freedom available in the channel

Cette thèse présente une étude des propriétés de la lumière émise par des diffuseurs atomiques distribués aléatoirement dans l'espace euclidien, et interagissant avec le champ électromagnétique. Dans ce cadre, une théorie ab initio des lasers aléatoires est formulée en terme des propriétés statistiques de la `matrice de Green'. Cette dernière appartient à la famille des matrices aléatoires euclidiennes (MAE) pour lesquelles nous développons une théorie analytique donnant notamment accès à la distribution de probabilité de leurs valeurs propres. Dans un premier temps, nous démontrons les équations quantiques microscopiques régissant la dynamique du champ électrique ainsi que celle des opérateurs atomiques, et explicitons comment la matrice de Green (dont les éléments sont égaux à la fonction de Green de l'équation de Helmholtz évaluée entre les différentes paires d'atomes constituant le milieu) émerge naturellement du formalisme quantique. Nous exprimons à la fois l'intensité et le spectre de la lumière en termes des propriétés de la matrice de Green, caractérisons les forces de Langevin quantiques, et montrons de quelle manière le seuil semi-classique d'un laser aléatoire est affecté par la prise en considération des fluctuations quantiques (chapitres 2 et 3). Une description mésoscopique et semi-classique de la lumière diffusée par un grand nombre d'atomes soumis à une pompe externe et distribués aléatoirement dans l'espace libre est présentée dans le quatrième chapitre. Après avoir établi une condition de seuil laser universelle, valide quelle que soit la configuration des atomes, nous démontrons une équation de transport obéie par l'intensité moyenne en présence de gain, discutons différentes approximations de cette dernière (équation de Bethe-Salpeter, équation de Boltzmann, équation de diffusion), établissons un `mapping' avec les MAE, et analysons la condition de seuil laser déduite de l'équation de transport. Poussés par la volonté de caractériser analytiquement les propriétés statistiques de la matrice de Green, nous développons dans les chapitres 5 et 6 une théorie générale des MAE, hermitiennes et non hermitiennes, valide dans la limite de grande taille matricielle. Nous obtenons des équations couplées pour la résolvante et le corrélateur des vecteur propres d'une MAE arbitraire, puis testons la validité de nos résultats sur trois matrices jouant un rôle important dans l'étude de la propagation des ondes en milieux désordonnés: la matrice de Green dans l'espace tridimensionnel, sa partie imaginaire, et sa partie réelle. D'un point de vue physique, nous sommes capables de décrire analytiquement avec une bonne précision la distribution de probabilité des taux d'émission lumineux dus à un grand nombre d'atomes, ainsi que celle du déplacement lumineux collectif dû à l'interaction lumière-matière. Par ailleurs, nous proposons d'utiliser la distribution des valeurs propres de la matrice de Green non hermitienne comme une carte unique sur laquelle peuvent s'identifier différents régimes de désordre (balistique, diffusif, localisé, milieu effectif, superradiance). Finalement, nous combinons les équations microscopiques de l'interaction lumière-matière avec nos résultats relatifs aux MAE non-hermitiennes afin de caractériser dans le détail le comportement des lasers aléatoires. Le seuil laser ainsi que l'intensité au delà du seuil sont calculés analytiquement dans l'approximation semi-classique, et le spectre de la lumière sous le seuil est évalué en prenant en compte les effets quantiques. Notre théorie s'applique aussi bien à basse densité qu'à haute densité de diffuseurs atomiques. / This thesis is devoted to the study of the properties of light emitted by a collection of atomic scatterers distributed at random positions in Euclidean space and interacting with the electromagnetic field. In this respect, an ab initio analytic theory of random lasing is formulated in terms of the statistical properties of the so-called `Green's matrix'. The latter belongs to the family of Euclidean random matrices (ERM's), for which we develop an analytic theory giving access to their eigenvalue distribution. First, we derive quantum microscopic equations for the electric field and atomic operators, and show how the non-Hermitian Green's matrix (a matrix with elements equal to the Green's function of the Hemholtz equation between pairs of atoms in the system) emerges in the quantum formalism. We provide expressions for the intensity and the spectrum of light in terms of the properties of the Green's matrix, characterize quantum Langevin forces, and reveal how the semiclassical random laser threshold is washed out by quantum fluctuations (chapters 2 and 3). A mesoscopic and semiclassical description of light scattered by an arbitrary large number of pumped atoms randomly distributed in free space is the subject of chapter 4. After deriving a universal lasing threshold condition valid for any configuration of atoms, we provide a microscopic derivation of transport equation in the presence of gain, discuss various approximations of the latter (Bethe-Salpeter, Boltzmann, diffusion equations), reveal a mapping to ERM's, and analyze the lasing threshold condition inferred from the transport equation. Facing the problem of characterizing analytically the statistical properties of the Green's matrix, we develop in chapters 5 and 6 a theory for Hermitian and non-Hermitian ERM's in the limit of large matrix size. We obtain self-consistent equations for the resolvent and the eigenvector correlator of arbitrary ERM and apply our results to three different ERM's relevant to wave propagation in random media: the three-dimensionnal Green's matrix, its imaginary part and its real part. From a physical point of view, we are able to describe analytically with a fair precision the full probability distribution of decay rates of light emitted by a large number of atoms, as well as of the collective frequency shift induced by the light-matter interaction. In addition, we promote the idea that the eigenvalue distribution of the Green's matrix can serve as a map on which signatures of various regimes of disorder can be distinguished (ballistic, diffusive, localized, effective medium, and superradiance regimes). Finally, we combine microscopic equations of motion of light-matter interaction with our results for non-Hermitian ERM's to tackle the problem of random lasing. Lasing threshold and the intensity of laser emission are calculated analytically in the semiclassical approximation, and the spectrum of light below threshold is computed by taking into account quantum effects. Our theory applies all the way from low to high density of atoms.

[pt] Em 1994, Henry Landau mostrou que uma matriz de Toeplitz real simétrica pode assumir qualquer valor real. O objetivo desse texto é apresentar a demonstração de Landau. São empregadas técnicas de teoria de grau topológico e teoria espectral. / [en] In 1994, Henry Landau proved that a real, symmetric Toeplitz matrix obtains an arbitrary real spectrum. In this text, we present the details of his proof. The key ingredients are topological degree theory and spectral theory.

Álgebra Linear e particularmente a teoria das matrizes e dos sistemas lineares são tópicos da Matemática que têm aplicações, não só dentro da própria Matemática, mas também em várias outras áreas do conhecimento humano. Neste trabalho, além de estudar estas teorias, estudamos algumas de suas aplicações na área da Economia, como em modelos lineares de produção, modelos de Markov para emprego e modelos de benefícios obtidos no pagamento de impostos após realizarmos contribuições filantrópicas. / Linear Algebra and particularly matrices and linear systems theory are topics in Mathematics with many applications in several branches of science. In this work we study this theory and some of its applications in Economy as in linear models of production, Markov models of employment and tax benefits of charitable contributions.