L’objectif de ces travaux de thèse est la conception et le développement d’un système de caractérisation tridimensionnelle de pièces forgées de grande dimension portées à haute température. Les travaux se basent sur de nombreuses thématiques telles que l’acquisition tridimensionnelle, l’extraction, la segmentation et le recalage de primitives 3D. Nous présentons tout d’abord les limites des systèmes de caractérisation de pièces forgées cités dans la littérature. Dans la deuxième partie, nous présentons la réalisation du système de caractérisation de pièces forgées, constitué de deux scanners temps de vol (TOF). Nous présentons également le simulateur de numérisation par scanner TOF qui nous permet de nous affranchir des contraintes industrielles (temps, difficulté de manœuvres) pour positionner les deux scanners. La troisième partie est consacrée à l’extraction des primitives 3D. Nous avons traité deux types de primitives : viroles et sphères avec deux approches différentes : méthode supervisée et méthode automatique. La première approche basée sur une méthode de croissance de région et de contour actif, permet d’extraire des formes extrudées complexes. Des problèmes d’ergonomie du système nous ont conduits à développer une deuxième approche, basée sur l’image de Gauss et l’extraction d’ellipse, qui permet l’extraction automatique de formes cylindriques ovales ou circulaires. Nous présentons également quatre méthodes d’extraction automatique de sphères basées sur des approches heuristiques : RANSAC (RANdom SAmple Consensus), algorithme génétique et algorithme génétique par niche. Dans la quatrième partie, nous étudions les différentes approches de recalage de données 3D traitées : le calibrage basé sur les cibles artificielles et le recalage fin basé sur l’algorithme ICP. Pour conclure, nous présentons la réalisation d’un système complet de caractérisation tridimensionnelle de pièces forgées de grande dimension. Ensuite, nous comparons les performances et les limites de ce système avec les systèmes de caractérisation cités dans la littérature. / The main objective of this Phd project is to conceive a machine vision system for hot cylindrical metallic shells diameters measurement during forging process. The manuscript is structured by developing in the first chapter the state of the art and the limits of hot metallic shells measurement systems suggested in literature. Our implemented system which is based on two conventional Time Of Flight (TOF) laser scanners has been described in the same chapter along, chapter two, with presentation of its respective numerical simulator. Simulation series have been done using the digitizing simulator and were aimed to determine the optimal positions of the two scanners without any industrial constraints (time, difficulty of operations). The third part of the manuscript copes with 3D primitives extraction. Two major types of approaches have been studied according to the primitive’s form (cylinders or spheres) to be extracted: supervised method and automatic method. The first approach, based on a growing region method and active contour, enables to extract complex extruded forms; while problems of ergonomics have been solved using automatic methods that have been carried out along the programme research. The proposed methods consist in automatically extracting: oval or circular cylindrical forms, using Gauss map associated with ellipse extraction techniques : spherical forms, using heuristic approaches such as RANdom SAmple Consensus RANSAC, Genetic Algorithm (GA) and Niche Genetic Algorithm (NGA). Two varieties of 3D data registration approach have been presented and discussed in chapter 4: the registration based on the artificial targets and the fine registration based on algorithm ICP. A complete system for three-dimensional characterization of hot cylindrical metallic shells during forging process has been implemented and then compared with existing systems in order to identify its performances and limits in conclusion.

Distribution network 110 kV owned by E. ON in the area Otrokovice; powered by 110 kV and two variants of involvement contained.The first option is basic involvement, without the use of the bridge. The second option includes involvement with the bridge. The aim of this study is to compare; by calculating the steady-state network operation and short circuit conditions of the network, the involvement of these two options. The thesis is divided into two parts, theoretical and practical. The theoretical part consists of a description of the steady operation of networks of high-voltage and short circuit ratio calculations. Load flow calculations are described by the Gauss-Seidel and Newton iterative method. In the case of short-circuit conditions, the effects of their characteristic values, processes and various methods of calculation are described.In the second part, this theoretical knowledge is applied to input data and dispatching programme with the appropriate calculations of network operation and short circuit conditions. The calculated values are listed in the thesis, on the basis of which an evaluation of the two possible connections is made.

In 1883 Leopold Kronecker published a paper containing “a few explanatory remarks” to an earlier paper of his from 1866. His work loosely connected the theory of integral binary bilinear forms to the theory of integral binary quadratic forms. In this dissertation we discover the statements within Kronecker's paper and offer detailed arithmetic proofs. We begin by developing the theory of binary bilinear forms and their automorphs, providing a classification of integral binary bilinear forms up to equivalence, proper equivalence and complete equivalence. In the second chapter we introduce the class number, proper class number and complete class number as well as two refinements, which facilitate the development of a connection with binary quadratic forms. Our third chapter is devoted to deriving several class number formulas in terms of divisors of the determinant. This chapter also contains lower bounds on the class number for bilinear forms and classifies when these bounds are attained. Lastly, we use the class number formulas to rigorously develop Kronecker's connection between binary bilinear forms and binary quadratic forms. We supply purely arithmetic proofs of five results stated but not proven in the original paper. We conclude by giving an application of this material to the number of representations of an integer as a sum of three squares and show the resulting formula is equivalent to the well-known result due to Gauss.

Thesis (PhD (Physics))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: There are many applications where a Gaussian laser beam is not ideal, for example, in areas such as medicine, data storage, science, manufacturing and so on, and yet in the vast majority of laser systems this is the fundamental output mode. Clearly this is a limitation, and is often overcome by adapting the application in mind to the available beam. A more desirable approach would be to create a laser beam as the output that is tailored for the application in mind – so called intra-cavity laser beam shaping. The main goal of intra-cavity beam shaping is the designing of laser cavities so that one can produce beams directly as the output of the cavity with the required phase and intensity distribution. Shaping the beam inside the cavity is more desirable than reshaping outside the cavity due to the introduction of additional external losses and adjustment problems. More elements are required outside the cavity which leads to additional costs and larger physical systems. In this thesis we present new methods for phase and amplitude intra– cavity beam shaping. To illustrate the methods we give both an analytical and numerical analysis of different resonator systems which are able to produce customised phase and intensity distributions. In the introduction of this thesis, a detailed overview of the key concepts of optical resonators is presented. In Chapter 2 we consider the well–known integral iteration algorithm for intra–cavity field simulation, namely the Fox–Li algorithm and a new method (matrix method), which is based on the Fox–Li algorithm and can decrease the computation time of both the Fox–Li algorithm and any integral iteration algorithms. The method can be used for any class of integral iteration algorithms which has the same calculation integrals, with changing integrants. The given method appreciably decreases the computation time of these algorithms and approaches that of a single iteration. In Chapter 3 a new approach to modeling the spatial intensity profile from Porro prism resonators is proposed based on rotating loss screens to mimic the apex losses of the prisms. A numerical model based on this approach is presented which correctly predicts the output transverse field distribution found experimentally from such resonators. In Chapter 4 we present a combination of both amplitude and phase shaping inside a cavity, namely the deployment of a suitable amplitude filter at the Fourier plane of a conventional resonator configuration with only spherical curvature optical elements, for the generation of Bessel–Gauss beams as the output. In Chapter 5 we present the analytical and numerical analyses of two new resonator systems for generating flat–top–like beams. Both approaches lead to closed form expressions for the required cavity optics, but differ substantially in the design technique, with the first based on reverse propagation of a flattened Gaussian beam, and the second a metamorphosis of a Gaussian into a flat–top beam. We show that both have good convergence properties, and result in the desired stable mode. In Chapter 6 we outline a resonator design that allows for the selection of a Gaussian mode by diffractive optical elements. This is made possible by the metamorphosis of a Gaussian beam into a flat–top beam during propagation from one end of the resonator to the other. By placing the gain medium at the flat–top beam end, it is possible to extract high energy in a low–loss cavity. / AFRIKAANSE OPSOMMING: Daar is verskeie toepassings waar ʼn Gaussiese laser bundel nie ideaal is nie, in gebiede soos mediese veld, stoor van data, vervaardiging en so meer, en tog word die meeste laser sisteme in die fundamentele mode bedryf. Dit is duidelik ’n beperking, en word meestal oorkom deur aanpassing van die toepassing tot die beskikbare bundel. ’n Beter benadering sou wees om ʼn laser bundel te maak wat afgestem is op die toepassing - sogenaamde intra-resonator bundel vorming. Die hoofdoel van intra-resonator bundel vorming is om resonators te ontwerp wat direk as uitset kan lewer wat die gewenste fase en intensiteits-distribusie vertoon. Vorming van die bundel in die resonator is voordeliger omdat die vorming buite die resonator tot addisionele verliese asook verstellings probleme bydra. Meer elemente word benodig buite die resonator wat bydra tot hoër koste en groter sisteme. In hierdie tesis word nuwe fase en amplitude intra-resonator bundelvormings metodes voorgestel. Om hierdie metode te demonstreer word analitiese en numeriese analises vir verskillende resonator sisteme wat aangepaste fase en intensiteit distribusies produseer, bespreek. In die inleiding van die tesis word ʼn detailleer oorsig oor die sleutel konsepte van optiese resonators voorgelê. In hoofstuk 2 word die bekende integraal iterasie algoritme vir intraresonator veld simulasie, naamlik die Fox-Li algoritme, en ʼn nuwe metode (matriks metode), wat gebaseer is op die Fox-Li algoritme, en die berekeningstyd van beide die Fox-Li algoritme en enige ander integraal iterasie algoritme verminder. Die metode kan gebruik word om enige klas van integraal iterasie algoritmes wat dieselfde berekenings integrale het, met veranderde integrante (waar die integrand die veld van die lig golf is in die geval van die Fox-Li algoritme, IFTA, en die skerm metode. Die voorgestelde metode verminder die berekeningstyd aansienlik, en is benaderd die van ʼn enkel iterasie berekening. In hoofstuk 3 word ʼn nuwe benadering om die modellering van die ruimtelike intensiteitsprofiel van Porro prisma resonators, gebaseer op roterende verliese skerms om die apeks-verliese van die prismas te benader, voorgestel. ʼn Numeriese model gebaseer op hierdie benadering wat die uitset van die transversale veld distribusie in eksperimentele resonators korrek voorspel, word voorgestel. In hoofstuk 4 word ʼn tegniek vir die generering van Bessel-Gauss bundels deur die gebruik van ʼn kombinasie van amplitude en fase vorming in die resonator en ʼn geskikte amplitude filter in die Fourier vlak van ʼn konvensionele resonator konfigurasie met optiese elemente wat slegs sferiese krommings het, voorgestel. In hoofstuk 5 word die analitiese en numeriese analises van twee nuwe resonator sisteme vir die generering van sogenaamde “flat–top” bundels voorgestel. Beide benaderings lei na ʼn geslote vorm uitdrukking vir die resonator optika wat benodig word, maar verskil noemenswaardig in die ontwerptegniek. Die eerste is baseer op die terug voortplanting van plat Gaussiese bundel, en die tweede op metamorfose van Gaussiese “flat-top” bundel. Ons toon aan dat beide tegnieke goeie konvergensie het, en in die gevraagde stabiele modus lewer. In hoofstuk 6 skets ons die resonator ontwerp wat die selektering van ʼn Gaussiese modus deur diffraktiewe optiese element moontlik maak. Dit word moontlik deur die metamorfose van ’n Gaussiese bundel na ʼn “flat-top” gedurende die voortplanting van die een kant van die resonator na die ander. Deur die wins medium aan die “flat–top” kant van die bundel te plaas word dit moontlik om hoë energie te onttrek in ʼn lae verlies resonator.

The successful application of Computational Aeroacoustics (CAA) requires high accuracy numerical schemes with good dissipation and dispersion characteristics. Unstructured meshes have a greater geometrical flexibility than existing high order structured mesh methods. This work investigates the suitability of unstructured mesh techniques by computing a two-dimensionallinearised Euler problem with various discretisation schemes and different mesh types. The goal of the present work is the development of an unstructured numerical method with the high accuracy, low dissipation and low dispersion required to be an effective tool in the study of aeroacoustics. The suitability of the unstructured method is investigated using aeroacoustic test cases taken from CAA Benchmark Workshop proceedings. Comparisons are made with exact solutions and a high order structured method. The higher order structured method was based upon a standard central differencing spatial discretisation. For the unstructured method a vertex-based data structure is employed. A median-dual control volume is used for the finite volume approximation with the option of using a Green-Gauss gradient approximation technique or a Least Squares approximation. The temporal discretisation used for both the structured and unstructured numerical methods is an explicit Runge-Kutta method with local timestepping. For the unstructured method, the gradient approximation technique is used to compute gradients at each vertex, these are then used to reconstruct the fluxes at the control volume faces. The unstructured mesh types used to evaluate the numerical method include semi-structured and purely unstructured triangular meshes. The semi-structured meshes were created directly from the associated structured mesh. The purely unstructured meshes were created using a commercial paving algorithm. The Least Squares method has the potential to allow high order reconstruction. Results show that a Weighted Least gradient approximation gives better solutions than unweighted and Green-Gauss gradient computation. The solutions are of acceptable accuracy on these problems with the absolute error of the unstructured method approaching that of a high order structured solution on an equivalent mesh for specific aeroacoustic scenarios.

En 1975 H. Fujimoto a généralisé les résultats d’unicité pour des fonctions holomorphes dus à Nevanlinna pour des applications méromorphes de Cn dans CPN. Il a démontré que pour deux applications méromorphes non linéairement dégénérées f et g de Cn dans CPN, si elles ont les mêmes images réciproques, comptées avec leurs multiplicités, par rapport à (3N + 2) hyperplans de CPN en position générale, alors f g. Depuis, ce problème a été étudié d’une manière intensive par H. Fujimoto, W. Stoll, L. Smiley, M. Ru, G. Dethloff-T.V.Tan, D.D.Thai-S.D.Quang, Chen-Yan et d’autres auteurs. En parallèle avec le développement de la théorie de Nevanlinna, la théorie de distribution des valeurs de l’application de Gauss des surfaces minimales dans Rm a été étudiée d’une manière intensive par R.Osserman, S.S. Chern, F. Xavier, H. Fujimoto, S.J. Kao, M. Ru et d’autres auteurs. Dans cette thèse, nous avons continué d’étudier ces problèmes. Nous avons obtenu les résultats principaux suivants: +) Théorèmes d’unicité avec multiplicités tronquées des applications méromorphes de Cn dans CPN ayant les mêmes images réciproques par rapport è (2N + 2) hyperplans de CPN. +) Théorèmes d’unicité avec multiplicités tronquées des applications méromorphes de Cn dans CPN ayant des cibles mobiles et un ensemble d’identité petit. +) Théorèmes d’unicité avec multiplicités tronquées des applications méromorphes de Cn dans CPN ayant des cibles fixes ou mobiles et satisfaisant des conditions sur les dérivées. +) Théorèmes de ramification de l’application de Gauss de certaines classes de surfaces minimales complètes dans Rm (m = 3,4). / In 1975, H. Fujimoto generalized Nevanlinna’s known results for meromorphic fonctions to the case of meromorphic mappings of Cn into PN(C). He proved that for two linearly nondegenerate meromorphic mappings f and g of C into PN(C). if they have the saine inverse images counted with multiplicities for 3N + 2 hyperplanes in general position in PN(C) then f = g. After that, this problem has been studied intensively by a number of mathematicans as H. Fujimoto, W. Stoll, L. Smiley, M. Ru, G. Dethloff - T. V. Tan, D. D. Thai - S. D. Quang, Chen-Yan and so on. Parallel to the development of Nevanlinna theory, the value distribution theory of the Gauss map of minimal surfaces immersed in Rm vas studied by many mathematicans as R. Osserman, S.S. Chern, F. Xavier, H. Fujimoto, S. J. Kao, M. Ru and many other mathematicans. In this thesis, we continuous studing some problems on these directions. The main goals of the thesis are followings. • Unicity theorems with truncated multiplicities of meromorphic mappings of Cn into PN(C) sharing 2N + 2 fixed hyperplanes.• Unicity theorems with truncated multiplicities of meromorphic mappings of Cn into PN(C) for moving targets, and a small set of identity.

Les modes de Laguerre-Gauss sont une des solutions de l’équation de propagation de la lumière dans l’approximation paraxiale en coordonnées cylindriques. Ces modes sont caractérisés par deux indices (azimutal et radial) et présentent des propriétés particulières : une structure en forme d’anneau et une phase en hélice. Ces propriétés ont été mises au profit pour plusieurs applications allant de la microscopie à l’astronomie. Cette thèse a été dédiée à l’étude de ces modes et à leur application dans le domaine des atomes froids. Dans un premier temps, la pureté des modes fabriqués par la méthode d’holographie numérique a été étudiée, ainsi que le rôle que la pureté joue dans leur propagation. Ces modes ont été ensuite utilisés pour l’obtention une source d’atomes froids brillante. Depuis plus de 20 ans, des efforts considérables ont été faits pour obtenir des sources d’atomes froids les plus brillantes et les plus compactes possibles. Un piège magnéto-optique à deux dimensions (2D-MOT) est couramment utilisé pour obtenir une source continue d’atomes froids avec un flux important de l’ordre de 1010 atomes/s. Toutefois, le jet atomique ainsi obtenu présente une divergence d’environ 40 mrad ce qui contraint l’utilisateur à travailler près de la sortie mais avec un accès optique limité ou bien loin où il a un bon accès optique mais au prix d’une densité atomique plus faible. Une alternative est présentée, dans laquelle un mode de Laguerre-Gauss est utilisé pour canaliser les atomes à la sortie d’un 2D-MOT. Le mode de Laguerre-Gauss réalise un piégeage dipolaire qui confine les atomes sont le centre noir du mode, ce qui présente l’intérêt de limiter le chauffage dû à l’absorption/émission de la lumière. Ainsi, puisque les modes de Laguerre-Gauss gardent leur forme au cours de leur propagation, les atomes sont canalisés sur une distance de plusieurs décimètres. Nous avons étudié le fonctionnement de ce système pour différents ordres du mode de Laguerre-Gauss et différentes fréquences, et nous avons montré que ce système est efficace et permet d’atteindre un gain en densité d’un facteur 200 par rapport au cas d’un 2D-MOT conventionnel. Un cas particulier est aussi présenté, dans lequel la fréquence du mode de Laguerre-Gauss est choisie pour effectuer non seulement la canalisation des atomes mais aussi leur repompage, ce qui conduit à une simplification importante du système. D’un autre côté, les modes de Laguerre-Gauss sont d’un grand intérêt dans le domaine de l’information et la cryptographie quantique car ils peuvent être utilisés pour encoder et enregistrer l’information. Pour cela, il est indispensable de pouvoir les détecter de façon non-équivoque. Dans ce contexte, nous nous sommes intéressés à la détermination de ces modes. Jusqu’à maintenant, les techniques de détection ont permis de mesurer l’indice azimutal mais peu d’entre elles mesurent l’indice radial. Si on est capable de mesurer aussi l’indice radial, il peut être utilisé comme une nouvelle variable pour le codage et l’enregistrement de l’information. Sous cette motivation, une technique basée sur la transformation du mode de Laguerre-Gauss par un système astigmatique a été mise au point pour déterminer les deux indices qui le caractérisent. Nous avons montré que cette technique peut aussi être utilisée pour optimiser expérimentalement la fabrication des modes de Laguerre-Gauss d’ordre élevé les plus purs possibles. / Laguerre-Gaussian modes of light are one of the solutions to the propagation equation of light in the paraxial approximation in cylindrical coordinates. These modes are characterized by two indices (azimuthal and radial) and present particular properties: ring-shaped structure and a helical phase. These properties have been put to use in several applications going from microscopy to astronomy. This work has been devoted to the study of these modes and their application in the cold atom domain. Initially, the purity of such modes generated by numerical holography was studied, as well as the role played by the purity in their propagation. These modes were then used to obtain a bright source of cold atoms. For more than 20 years, considerable efforts have been made to create sources of cold atoms as bright and compact as possible. A two dimensional magneto-optical trap (2D-MOT) is currently used to obtain a continuous source of cold atoms with high flux of the order of 1010 atomes/s. Nevertheless, the source of cold atoms thereby achieved show a divergence of about 40 mrad which constrains the user to work by output but with a limited optical access or far from the output but at the cost of lower density. An alternative is presented, in which a Laguerre-Gaussian mode of light is used to channel the atoms at the output of a 2D-MOT. The Laguerre-Gaussian mode of light carries out a dipole trapping that confines the atoms in the black center of the mode, which has the advantage of limiting the heating due to absorption/emission of light. Thus, since the Laguerre-Gaussian modes of light keep their ring-shape along their propagation, the atoms are channeled over a distance of several decimeters. We studied the functioning of this system for different orders of the Laguerre-Gaussian mode and at different frequencies, and we showed that this system is efficient and allows to achieve a density gain of a factor 200 compared to a conventional 2D-MOT. A particular case is also presented, in which the frequency of the Laguerre-Gaussian mode of light is chosen to carry out not only the channeling of the atoms, but also their repumping, which leads to an important simplification of the system. Laguerre-Gaussian modes of light are also of great interest in the domain of quantum information and quantum cryptography because they can be used to encode and store information. In order to do that, it is crucial to be able to detect them in an unequivocal way. In this context, we are interested in the determination of these modes. Until now, different detection techniques have measured the azimuthal index, but few of them have been able to measure the radial index. If we are able to measure the radial index, it can be used as a new variable to encode and store information. Under this motivation, a technique based on the transformation of a Laguerre-Gaussian mode of light by an astigmatic system has been extended to determine the two indices characterizing them. We showed that this technique can also be used to experimentally optimize the generation of the purest possible Laguerre-Gaussian modes of high order.

Bundle Adjustment is a common fine-tuning step used in photogrammetry. It uses different types of parameters, some of which can be considered to be almost linear while others can be considered to be highly nonlinear, e.g. the rotational parameters. However, in the Bundle Adjustment process all parameters are treated equal. In concert with a poor initial estimate, this might cause Bundle Adjustment to diverge. In this report, two novel methods based on the damped Gauss-Newton with Armijo linesearch, modified by giving rotational parameters a special treatment, are tested. These methods, Clamped Alpha and Linear Exponential Search, are compared to Gauss-Newton with Armijo linesearch, as well as to the undamped Gauss-Newton method, also known as the Gauss-Markov method. Parameter sweeps over different perturbation levels for the angular parameters show that each of the three damped methods outperform the Gauss-Newton method. Notably, the Clamped Alpha method also outperforms the other two damped methods, with as much as 16 times as many convergent cases for a given perturbation level. Meanwhile, the average number of iterations is increased by only 1.8 times that of the Gauss-Newton with Armijo linesearch. The results add to existing research arguing for the use of damped methods in Bundle Adjustment. In particular, the simple and cheap Clamped Alpha method is potentially attractive for problems where the uncertainty of the camera angles is significant. While the Clamped Alpha method show promising results, it should be noted that the experiments in this study are on synthetic data. In order to solidify these results, further investigations into the performance of Clamped Alpha using real-world data should be conducted.

A Tomografia por Impedância elétrica (TIE) tem como objetivo gerar imagens da distribuição de resistividade dentro de um domínio. A TIE injeta correntes em eletrodos alocados na fronteira do domínio e mede potenciais elétricos através dos mesmos eletrodos. A TIE é considerada um problema inverso, não-linear e mal posto. Atualmente, para gerar uma solução do problema inverso, existem duas classes de algoritmos para estimar a distribuição de resistividade no interior do domínio, os que estimam variações da distribuição de resistividade do domínio e os absolutos, que estimam a distribuição de resistividade. Variações da distribuição de resistividade são o resultado da solução de um sistema linear do tipo Ax = b. O objetivo do presente trabalho é avaliar o desempenho da Programação Linear (PL) na solução do sistema linear, avaliar o algoritmo quanto a propaga- ção de erros numéricos e avaliar os efeitos de restringir o espaço solução através de restrições de PL. Os efeitos do uso de Programação Linear é avaliado tanto em métodos que geram imagens de diferenças, como o Matriz de Sensibilidade, como em métodos absolutos, como o Gauss-Newton. Mostra-se neste trabalho que o uso da PL diminui o erro numérico propagado quando comparado ao uso do algoritmo LU Decomposition. Resulta também que reduzir o espaço solução, diretamente através de restrições de PL, melhora a resolução em resistividade e a resolução espacial da imagem quando comparado com o uso de LU Decomposition. / Electrical impedance tomography (EIT) generates images of the resistivity distribution of a domain. The EIT method inject currents through electrodes placed on the boundary of the domain and measures electric potentials through the same electrodes. EIT is considered an inverse problem, non-linear and ill-conditioned. There are two classes of algorithms to estimate the resistivity distribution inside the domain, difference images algorithms, which estimate resistivity distribution variations, and absolute images algorithms, which estimate the resistivity distribution. Resistivity distribution variations are the solution of a linear system, say Ax = b. In this work, the main objective is to evaluate the performance of Linear Programming (LP) solving an EIT linear system from the point of view of the numerical error propagation and the ability to constrain the solution space. The impact of using LP to solve an EIT linear system is evaluated on a difference image algorithm and on an absolute algorithm. This work shows that the use of LP diminishes the numerical error propagation compared to LU Decomposition. It is also shown that constraining the solution space through LP improves the resistivity resolution and the spatial resolution of the images when compared to LU Decomposition.

Differential-algebraic equations (DAEs) consist of mixed systems of ordinary differential equations (ODEs) coupled with linear or nonlinear equations. Such systems may be viewed as ODEs with integral curves lying in a manifold. DAEs appear frequently in applications such as classical mechanics and electrical circuits. This thesis concentrates on systems of index 2, originally index 3, and mixed index 2 and 3. Fast and efficient numerical solvers for DAEs are highly desirable for finding solutions. We focus primarily on the class of Gauss-Lobatto SPARK methods. However, we also introduce an extension to methods proposed by Murua for solving index 2 systems to systems of mixed index 2 and 3. An analysis of these methods is also presented in this thesis. We examine the existence and uniqueness of the proposed numerical solutions, the influence of perturbations, and the local error and global convergence of the methods. When applied to index 2 DAEs, SPARK methods are shown to be equivalent to a class of collocation type methods. When applied to originally index 3 and mixed index 2 and 3 DAEs, they are equivalent to a class of discontinuous collocation methods. Using these equivalences, (s,s)--Gauss-Lobatto SPARK methods can be shown to be superconvergent of order 2s. Symplectic SPARK methods applied to Hamiltonian systems with holonomic constraints preserve well the total energy of the system. This follows from a backward error analysis approach. SPARK methods and our proposed EMPRK methods are shown to be Lagrange-d'Alembert integrators. This thesis also presents some numerical results for Gauss-Lobatto SPARK and EMPRK methods. A few problems from mechanics are considered.