Doctor of Philosophy / Department of Mechanical and Nuclear Engineering / Steven J. Eckels / The current research presents the experimental investigation of the effect of tube pitch on enhanced tube bundles’ performance. The typical application of this research is flooded refrigerant evaporators. Boosting evaporator’s performance through optimizing tube spacing reduces cost and energy consumption. R-134a with the enhanced tube Turbo BII-HP and R-123 with Turbo BII-LP were used in this study. Three tube pitches were tested P/D 1.167, P/D 1.33, and P/D 1.5. Each tube bundle includes 20 tubes (19.05 mm outer diameter and 1 m long each) constructed in four passes. The test facility’s design allows controlling three variables, heat flux, mass flux, and inlet quality. The type of analysis used is local to one location in the bundle. This was accomplished by measuring the water temperature drop in the four passes. The water-side pressure drop is included in the data analysis. A new method called the EBHT (Enthalpy Based Heat Transfer) was introduced, which uses the water-side pressure drop in performing the heat transfer analysis. The input variables ranges are: 15-55 kg/m².s for mass flux, 5-60 kW/m² for heat flux, and 10-70% for inlet quality. The effect of local heat flux, local quality, and mass flux on the local heat transfer coefficient was investigated. The comparison between the bundle performance and single tube performance was included in the results of each tube bundle. The smallest tube pitch has the lowest performance in both refrigerants, with a significantly lower performance in the case of R-134a. However, the two bigger tube pitches have very similar performance at low heat flux. Moreover, the largest tube pitch performance approaches that of the single tube at medium and high heat fluxes. For the R-123 study, the smallest tube bundle experienced quick decease in performance at high qualities, exhibiting tube enhancement dry-out at certain flow rates and high qualities. The flow pattern effect was demonstrated by the dry-out phenomena. At medium and high heat fluxes, as the tube pitch increases, the performance approaches that of the single tube. All tube bundles experience quick decrease in performance at high qualities. Evidently, P/D 1.33 is the optimum tube pitch for the studied refrigerants and enhanced tubes combinations.

Nous étudions dans cette thèse le roulement sans glissement et sans pivotement de deux variétés lisses M et Ṁ l'une sur l'autre de dimensions et n et ṅ respectivement. L'objectif principal est de chercher des conditions nécessaires et suffisantes de la commandabilité du système commandé défini par le roulement. Dans le premier chapitre, on présente les motivations et le plan de la thèse ainsi les notations utilisées le long des chapitres. Dans le deuxième chapitre, on caractérise l'espace d'état du roulement quand M et Ṁ sont des variétés Riemanniennes lorsque n n'est pas nécessairement égal à ṅ et du développement quand M et Ṁ sont des variétés affines munies des connexions affines avec n = ṅ Ainsi, on donne les relèvements et les distributions correspondant aux deux notions précédentes. Le troisième chapitre contient quelques résultats de la commandabilité du système de roulement des variétés Riemanniennes. Plus précisément, on présente les conditions nécessaires de la non-commandabilité du roulement d'une variété Riemannienne 3-dimensionnelle sur une autre 2-dimensionnelle.Le chapitre 4 porte sur le roulement d'une variété Riemannienne de dimension 2 sur une autre de dimension 3. On trouve que la dimension d'une orbite non-ouverte quelconque de l'espace d'état appartient à {2,5,6,7}. Les aspects géométriques de deux variétés sont liés principalement avec le fait que la variété de dimension 3 contient une sous-variété totalement géodésique de dimension 2.Dans le dernier chapitre, on introduit et étudie un concept d'holonomie horizontale associé à un triplet (M,∇,Δ ) avec M variété différentielle connexe, ∇ connection affine complète sur M et Δ distribution complètement commandable. Si H^∇est le groupe d'holonomie associé à Ṁ on considère alors son sous-groupe obtenu uniquement en considérant le transport ∇- parallèle par rapport aux lacets dans M tangents à la distribution Δ On le note H_Δ^∇et on l’appelle groupe d'holonomie horizontal. On prouve que le groupe d'holonomie horizontal H_Δ^∇ est un sous-groupe de Lie de GL(n). Puis, on démontre par un exemple que la fermeture du groupe d'holonomie horizontal restreint (H_Δ^∇ )^0 n'est pas nécessairement égal à H_Δ^∇. A cette fin, on utilise le modèle du roulement avec M un groupe de Carnot homogène munie d'une connexion de Levi-Civita associée à une métrique Riemannienne sur l'espace Euclidien R^n munie de la connexion Euclidienne. / In this thesis, we study the rolling motion without spinning nor slipping of a smooth manifolds M and Ṁ against another of dimensions n and ṅ respectively. The purpose is to find the necessary and sufficient conditions for the controllability issue of the system of rolling. We start by a French review of the principal results of the thesis is included in the introduction. In Chapter 1, we present the motivations of the subject thesis, the structure of the contents and the notations used along the manuscript. The second chapter contain a characterization of the state space of rolling manifolds when M and Ṁ are Riemannian manifolds with n and ṅ are not necessarily equal and of the development of manifolds when M and Ṁ are affine manifolds of dimension n = ṅ equipped with affine connections. We also state the definitions of the lifts and the distributions with respect to the previous notions. The controllability results of the rolling system of Riemannian manifolds is included in Chapter 3. We give all the necessary conditions of the non-controllability of rolling of 3-dimensional Riemannian manifold against 2-dimensional Riemannian manifold. Chapter 4 deals with the rolling of a 2-dimensional Riemannian manifold against a 3-dimensional Riemannian manifold. We prove that the dimension of an arbitrary non-open orbit of the state space belongs to {2,5,6,7}. The geometrical aspects of the two manifolds depend on the existence of a 2-dimensional totally geodesic submanifold in the 3-dimensional manifold. The last chapter introduces and addresses the issue of horizontal holonomy associated to a triple (M,∇,Δ) with M smooth connected manifold, ∇ complete affine connection M and Δ completely controlable distribution over M. If H_Δ^∇. denotes the holonomy group associated with (M,∇) one considers its subgroup obtained by considering only the ∇- parallel transport with respect to loops of M tangent to the distribution Δ This subgroup is denoted by H_Δ^∇ and we call it horizontal holonomy group. We prove that the horizontal holonomy group H_Δ^∇ is a Lie subgroup of GL(n). Then, we show by means of an example that the closure of a restricted horizontal holonomy group on a Riemannian manifold is not necessarily equal to the holonomy group of the Riemannian manifold. To this end, we use the rolling problem of M taken as a step 2 homogeneous Carnot group equipped with the Levi-Civita connection associated to a Riemannian metric onto the Euclidean space R^n equipped with the Euclidean connection.

Introdução: a cintilografia cardíaca com MIBI-99mTc sincronizada ao eletrocardiograma (gated SPECT) avalia integridade celular, perfusão miocárdica e função global e regional do ventrículo esquerdo (VE). A terapia de ressincronização cardíaca (TRC) pode melhorar os sintomas de insuficiência cardíaca (IC), mas seus benefícios sobre a função do VE são menos pronunciados. Objetivos: avaliar se as mudanças na captação miocárdica do MIBI-99mTc após a TRC estão associadas à melhora clínica, à redução do QRS ao eletrocardiograma e ao desempenho do VE e se a gated SPECT adiciona informação na seleção e acompanhamento de pacientes para a TRC. Método: trinta pacientes (idade media 59 ± 11 anos, 47% masculinos) com miocardiopatia dilatada não isquêmica, IC classe funcional III ou IV da New York Heart Association com tratamento medicamentoso otimizado, bloqueio de ramo esquerdo e fração de ejeção do ventrículo esquerdo (FEVE) inferior a 35% participaram deste estudo. Foram avaliados pré e três meses após TRC as seguintes variáveis: classe funcional de IC, duração do QRS, FEVE pela ecocardiografia, captação miocárdica do MIBI-99mTc (%) ao repouso e após estresse com adenosina, volumes diastólico (VDF) e sistólico finais (VSF) do VE, motilidade e espessamento regionais nas paredes do VE pela Gated SPECT. Após TRC, os pacientes foram divididos em dois grupos de acordo com a melhora na FEVE: grupo 1 (G1=12 pacientes) com aumento 5 pontos absolutos; e grupo 2 (G2=18 pacientes) sem aumento significante. Resultados: após TRC, ambos os grupos melhoraram significantemente a classe funcional de IC, reduziram QRS e aumentaram a captação miocárdica do MIBI-99mTc nas paredes ântero-septal e ínfero-septal. Apenas G1 apresentou mudanças favoráveis no VDF, VSF, motilidade e espessamento regionais do VE. Pré TRC, pela análise univariada, o VDF e ESV foram estatisticamente maiores no G2 em relação ao G1 (VDF: 477 ± 168 mL vs. 276 ± 94 mL, p<0,001; VSF: 401 ± 154 mL vs. 220 ± 85 mL, p<0,001, G2 e G1, respectivamente). A captação miocárdica do MIBI-99mTc foi menor no G2 em relação ao G1 nas paredes anterior (60 ± 10% vs. 67 ± 7%, p=0,049, repouso) e inferior (48 ± 10% vs. 59 ± 11%, repouso, e 47 ± 10% vs. 58 ± 9%, p=0,003, após estresse). A soma dos escores de hipocaptação após estresse foi significantemente maior no G2 em relação ao G1 (14 ± 9 e 9 ± 4, G2 e G1, respectivamente, p=0,039). Pela análise multivariada, o VDF foi o único preditor independente de aumento na FEVE após terapia, p=0,01. O ótimo ponto de corte do VDF pela curva ROC para predizer melhora na FEVE após terapia foi 315 mL com sensibilidade de 89% e especificidade de 94%. Conclusões: A TRC aumentou a captação miocárdica regional de MIBI-99mTc, melhorou a classe funcional de IC e reduziu QRS independentemente da melhora do desempenho cardíaco. Após TRC, o aumento da FEVE ocorreu em corações menos dilatados e com uma maior captação miocárdica regional do MIBI-99mTc, principalmente na parede inferior / Background: gated myocardial scintigraphy with 99mTc-MIBI (gated SPECT) evaluates myocyte integrity, myocardial perfusion and global and regional left ventricular (LV) function. Cardiac resynchronization therapy (CRT) may improve heart failure (HF) clinical symptoms, but its benefits for LV function are less pronounced. Objectives: we assessed whether changes on myocardial 99mTc-MIBI uptake post-CRT are related to improvement in clinical HF symptoms, reduction in QRS duration and improvement in LV performance, and whether gated SPECT with 99mTc-MIBI adds information to selection and follow-up of patients undergoing CRT. Methods: thirty patients (mean age 59 ± 11, 47% male) with nonischemic dilated cardiomyopathy, chronic HF in New York Heart Association functional class III or IV with optimized clinical treatment, left bundle-branch block, and LV ejection fraction (LVEF) inferior to 35% participated to this study. Pre- and 3 months post-CRT, the variables analyzed were: HF functional class, QRS duration, LVEF by echocardiography, myocardial 99mTc-MIBI uptake (%) at rest and after adenosine stress, LV end-diastolic (EDV) and end-systolic volumes (ESV), regional LV motion and thickness by gated SPECT. Post-CRT, patients were divided into 2 groups according to LVEF improvement: group 1 (G1=12 patients) with increase in LVEF5points and group 2 (G2=18 patients) without significant increase. Results: post-CRT, both groups improved significantly in HF functional class, reduced QRS width and increased antero-septal and infero-septal wall 99mTc-MIBI uptake. Only G1 had favorable changes in EDV, ESV, and LV regional motion and thickness. Pre- CRT, by univariate analysis, EDV and ESV were statistically higher in G2 compared with G1 (EDV: 477 ± 168 mL vs. 276 ± 94 mL, p<0.001; ESV: 401 ± 154 mL vs. 220 ± 85 mL, p<0.001, G2 and G1, respectively). Myocardial 99mTc-MIBI uptake was lower in G2 compared with G1 in the anterior wall (60 ± 10% vs. 67 ± 7%, p=0.049, at rest) and inferior wall (48 ± 10% vs. 59 ± 11%, at rest, and 47 ± 10% vs. 58 ± 9%, p=0.003, after adenosine stress). Summed stress score was significantly higher in G2 compared to G1 (14 ± 9 vs. 9 ± 4, G2 and G1, respectively, p=0.039). By multivariate analysis, EDV was the only independent predictor of LVEF increase posttherapy, p=0.01. By ROC curve, optimal EDV cutoff point was 315mL with 89% of sensitivity and 94% of specificity. Conclusions: CRT increased myocardial 99mTc- MIBI uptake, improved HF functional class, and reduced QRS width independently of LV performance improvement. Post-CRT, LVEF increase occurred in hearts less dilated showing higher regional myocardial 99mTc-MIBI uptake, mainly in the inferior wall

La variabilité inter-individuelle en localisation auditive a été étudiée à travers plusieurs expérimentations. La première expérimentation a exploré la variabilité inter-individuelle en localisation auditive de sources sonores virtuelles (HRTFs individuelles et non individuelles) dans une population de 25 auditeurs naïfs préalablement soumis à un entraînement procédural. D'après nos résultats, la variabilité inter-individuelle dans la dimension gauche/droite pourrait être liée à une préférence individuelle vers un type d'indices parmi ceux utilisables pour la localisation dans cette dimension (indices binauraux/spectraux). Dans les dimensions haut/bas et avant/arrière, la variabilité inter-individuelle pourrait être liée à l'attention spatiale et à sa variation entre les zones de l'espace. Une seconde expérimentation réalisée chez les mêmes auditeurs a montré qu'une amélioration des capacités de localisation était possible par un entraînement par feedback visuel ou audio-moteur. Cette amélioration était souvent fonction de la performance avant entraînement, conduisant ainsi à une réduction de la variabilité inter-individuelle après entraînement. La variabilité inter-individuelle dans la dimension haut/bas après entraînement était en partie expliquée par l'activité du faisceau olivocochléaire efférent médian (FOCEM), structure du système auditif probablement impliquée dans le codage des indices spectraux de localisation. Ce résultat a été conforté par les résultats d'une troisième expérimentation qui ont révélé que les auditeurs dont la performance de localisation était la moins altérée dans le bruit étaient ceux dont le FOCEM avait été déterminé comme le plus actif. / Several experiments were performed to investigate interindividual variability in auditory localization. The first experiment explored interindividual variability in the localization of virtual sound sources (individualized and non-individualized HRTFs) in 25 naive listeners following procedural training. The results suggest that interindividual variability in localization performance in the left-right dimension stems from interindividual differences in the use of the different types of cues available for sound localization in this dimension (binaural/spectral cues). Interindividual variability in sound-localization performance in the up-down and front-back dimensions appears to be related primarily to spatial attention and to its variation across the area of space. A second experiment in the same group of listeners showed that localization performance can be improved by the provision of visual or auditory-motor feedback. The improvement depended upon pre-training performance, in such a way that interindividual variability in performance was smaller after training than before training. In addition, interindividual variability in the up-down dimension after training was found to be related to interindividual differences in the activity of the medial olivocochlear bundle (MOCB), an efferent auditory system that is currently likely thought to be involved in the processing of spectral cues for sound localization. This result was supported by the results of a third experiment which revealed that the listeners in whom localization performance was least impacted by background noise were those in whom the strongest MOCB had been measured.

Introduction : Le système nerveux autonome (SNA) est en situation supralévatorienne, le système somatique en situation infra lévatorienne. Le sympathique assurerait les fonctions de sécrétion et le parasympahtique, les fonctions dʼérection. Le rectum est en rapport étroit avec ces éléments nerveux. La protectomie avec exérèse totale du mésorectum pour cancer est associée à des séquelles sexuelles par lésion iatrogène des nerfs pelviens.Objectifs : Étudier la physiologie et lʼanatomie topographique et structurelle de l'innervation pelvipérinéale Matériels et méthodes : Cinq pelvis de foetus et dix pelvis dʼadultes féminins ont été prélevés pour études macroscopiques, microscopiques etimmunohistochimiques des nerfs pelviens. Les coupes ont été colorées puis immunomarquées pour détecter les fibres nerveuses (PS-100), somatiques(PMP22), autonomes adrénergiques (TH), cholinergiques (VAChT), sensitives(CGRP) et les fibres pro-érectiles (nNOS). Les lames ont été numérisées et reconstruites en 3D.Résultats : Les fibres nerveuses du SNA, richement interconnectées,véhiculent de façon mixte lʼinflux sympathique et parasympathique.Celles issues du plexus hypogastrique inférieur contrôlant les fonctions sexuelles sont regroupées avec le pédicule vaginal long et forment la bandelette neuro-vasculaire (BNV) sur la face antérolatérale du rectum à « 2et 10h ». Cette BNV est en avant de lʼexpansion postérolatérale du septum recto vaginal (SRV) qui la protège et qui est lʼéquivalent chez lʼhomme du fascia de Denonvilliers. Cette BNV concentre lʼensemble des fibres nerveusespro-érectiles destinées au périnée. Une lésion tronculaire de cette bandelette pourrait conduire à un trouble de lʼérection clitoridienne et de la lubrification vulvaire car, à ce niveau, les efférences sympathiques et parasympathiques coexistent.Conclusion: Un modèle anatomo-physiologique et pédagogiquedʼinnervation pelvipérinéale féminine a été développé. Ces travaux offrent des perspectives dʼétudes cliniques afin de mieux évaluer les dysfonctions sexuelles postopératoires. / Introduction: The nerve supply of the autonomic nervous system (ANS) to the pelvis is located above the levator ani muscle, and the somatic nerve supply to the pelvis is situated below levator ani. Sympathetic innervation assures secretory functions and parasympathetic innervation allows erection. The rectum is anatomically closely associated with these nerves. Protectomy with total excision of the mesorectum for cancer is associated with sexual sequellae due to iatrogenic damage to the pelvic nerves.Objectives: To study the physiology and topographic and structural anatomy of the innervation of the pelvis.Materials and methods: Five fœtal pelvises and ten adult female pelvises were collected for macroscopic, microscopic, and immunohistochemical studies of pelvic nerves. Sections were stained and then immunostained to reveal nerve fibres (PS-100), somatic nerves (PMP22), adrenergic autonomic nerves (TH), cholinergic autonomic nerves (VAChT), sensory nerves (CGRP) and pro-erectile nerves (nNOS). Sections were numbered and reconstructed in 3D.Results: ANS nerve fibres, densely interconnected, carry a combination of sympathetic and parasympathetic fibres.Nerve fibres controlling sexual function from the inferior hypogastric plexus are clustered along the vaginal pedicle and form the neurovascular bundle (NVB) on the anterolateral face of the rectum between “2 and 10 o’clock”. This NVB is in front of the expansion of the rectovaginal septum (RVS) which protects it. In males, the equivalent structure is the rectoprostratic fascia. This NVB contains all of the pro-erectile nerves supplying the perineum. A truncal lesion to this bundle could result in erectile dysfunction of the clitoris as well as difficulties in vulvar lubrication because sympathetic and parasympathetic efferent fibres are both present at this site.Conclusion: An educational anatomical and physiological model of the innervation of the female pelvis and perineum has been developed. This work offers perspectives for clinical studies to facilitate better evaluation of cases of post-operative sexual dysfunction.

Les travaux présentés dans ce manuscrit concernent le domaine de la localisation et la reconstruction 3D par vision artificielle. Dans ce contexte, la trajectoire d’une caméra et la structure3D de la scène filmée sont initialement estimées par des algorithmes linéaires puis optimisées par un algorithme non-linéaire, l’ajustement de faisceaux. Cette thèse présente tout d’abord une technique de recherche de l’amplitude de déplacement (recherche linéaire), ou line search pour les algorithmes de minimisation itérative. La technique proposée est non itérative et peut être rapidement implantée dans un ajustement de faisceaux traditionnel. Cette technique appelée recherche linéaire algébrique globale (G-ALS), ainsi que sa variante à deux dimensions (Two way-ALS), accélèrent la convergence de l’algorithme d’ajustement de faisceaux. L’approximation de l’erreur de reprojection par une distance algébrique rend possible le calcul analytique d’une amplitude de déplacement efficace (ou de deux pour la variante Two way-ALS), par la résolution d’un polynôme de degré 3 (G-ALS) ou 5 (Two way-ALS). Nos expérimentations sur des données simulées et réelles montrent que cette amplitude, optimale en distance algébrique, est performante en distance euclidienne, et permet de réduire le temps de convergence des minimisations. Une difficulté des algorithmes de localisation en temps réel par la vision (SLAM monoculaire) est que la trajectoire estimée est souvent affectée par des dérives : dérives d’orientation, de position et d’échelle. Puisque ces algorithmes sont incrémentaux, les erreurs et approximations sont cumulées tout au long de la trajectoire, et une dérive se forme sur la localisation globale. De plus, un système de localisation par vision peut toujours être ébloui ou utilisé dans des conditions qui ne permettent plus temporairement de calculer la localisation du système. Pour résoudre ces problèmes, nous proposons d’utiliser un capteur supplémentaire mesurant les déplacements de la caméra. Le type de capteur utilisé varie suivant l’application ciblée (un odomètre pour la localisation d’un véhicule, une centrale inertielle légère ou un système de navigation à guidage inertiel pour localiser une personne). Notre approche consiste à intégrer ces informations complémentaires directement dans l’ajustement de faisceaux, en ajoutant un terme de contrainte pondéré dans la fonction de coût. Nous évaluons trois méthodes permettant de sélectionner dynamiquement le coefficient de pondération et montrons que ces méthodes peuvent être employées dans un SLAM multi-capteur temps réel, avec différents types de contrainte, sur l’orientation ou sur la norme du déplacement de la caméra. La méthode est applicable pour tout autre terme de moindres carrés. Les expérimentations menées sur des séquences vidéo réelles montrent que cette technique d’ajustement de faisceaux contraint réduit les dérives observées avec les algorithmes de vision classiques. Ils améliorent ainsi la précision de la localisation globale du système. / The works presented in this manuscript are in the field of computer vision, and tackle the problem of real-time vision based localization and 3D reconstruction. In this context, the trajectory of a camera and the 3D structure of the filmed scene are initially estimated by linear algorithms and then optimized by a nonlinear algorithm, bundle adjustment. The thesis first presents a new technique of line search, dedicated to the nonlinear minimization algorithms used in Structure-from-Motion. The proposed technique is not iterative and can be quickly installed in traditional bundle adjustment frameworks. This technique, called Global Algebraic Line Search (G-ALS), and its two-dimensional variant (Two way-ALS), accelerate the convergence of the bundle adjustment algorithm. The approximation of the reprojection error by an algebraic distance enables the analytical calculation of an effective displacement amplitude (or two amplitudes for the Two way-ALS variant) by solving a degree 3 (G-ALS) or 5 (Two way-ALS) polynomial. Our experiments, conducted on simulated and real data, show that this amplitude, which is optimal for the algebraic distance, is also efficient for the Euclidean distance and reduces the convergence time of minimizations. One difficulty of real-time tracking algorithms (monocular SLAM) is that the estimated trajectory is often affected by drifts : on the absolute orientation, position and scale. Since these algorithms are incremental, errors and approximations are accumulated throughout the trajectory and cause global drifts. In addition, a tracking vision system can always be dazzled or used under conditions which prevented temporarily to calculate the location of the system. To solve these problems, we propose to use an additional sensor measuring the displacement of the camera. The type of sensor used will vary depending on the targeted application (an odometer for a vehicle, a lightweight inertial navigation system for a person). We propose to integrate this additional information directly into an extended bundle adjustment, by adding a constraint term in the weighted cost function. We evaluate three methods (based on machine learning or regularization) that dynamically select the weight associated to the constraint and show that these methods can be used in a real time multi-sensor SLAM, and validate them with different types of constraint on the orientation or on the scale. Experiments conducted on real video sequences show that this technique of constrained bundle adjustment reduces the drifts observed with the classical vision algorithms and improves the global accuracy of the positioning system.

Uma maneira de resolver problemas de programação linear de grande escala é explorar a relaxação lagrangeana das restrições \"difíceis\'\' e utilizar métodos de subgradientes. Populares por fornecerem rapidamente boas aproximações de soluções duais, eles não produzem diretamente as soluções primais. Para obtê-las com custo computacional adequado, pode-se construir seqüências ergódicas ou utilizar uma técnica proposta recentemente, denominada algoritmo do volume. As propriedades teóricas de convergência não foram bem estabelecidas nesse algoritmo, mas pequenas modificações permitem a demonstração da convergência dual. Destacam-se como adaptações o algoritmo do volume revisado, um método de feixes específico, e o algoritmo do volume incorporado ao método de variação do alvo. Este trabalho foi baseado no estudo desses algoritmos e de todos os conceitos envolvidos, em especial, análise convexa e otimização não diferenciável. Estudamos as principais diferenças teóricas desses métodos e realizamos comparações numéricas com problemas lineares e lineares inteiros, em particular, o corte máximo em grafos. / One way to solve large-scale linear programming problems is to exploit the Lagrangian relaxation of the difficult constraints and use subgradient methods. Such methods are popular as they give good approximations of dual solutions. Unfortunately, they do not directly yield primal solutions. Two alternatives to obtain primal solutions under reasonable computational cost are the construction of ergodic sequences and the use of the recently developed volume algorithm. While the convergence of ergodic sequences is well understood, the convergence properties of the volume algorithm is not well established in the original paper. This lead to some modifications of the original method to ease the proof of dual convergence. Three alternatives are the revised volume algorithm, a special case of the bundle method, and the volume algorithm incorporated by the variable target value method. The aim of this work is to study such algorithms and all related concepts, especially convex analysis and nondifferentiable optimization. We analysed the main theoretical differences among the methods and performed numerical experiments with linear and integer problems, in particular, the maximum cut problem on graphs.

Submitted by Bruno Caldeira Carlotti de Souza null (brunoccarlotti@gmail.com) on 2017-03-02T01:45:21Z No. of bitstreams: 1 Dissertação - Bruno C. C. de Souza.pdf: 1030573 bytes, checksum: e3dd1e43953565236359b6d10831067c (MD5) / Approved for entry into archive by LUIZA DE MENEZES ROMANETTO (luizamenezes@reitoria.unesp.br) on 2017-03-07T18:32:31Z (GMT) No. of bitstreams: 1 souza_bcc_me_sjrp.pdf: 1030573 bytes, checksum: e3dd1e43953565236359b6d10831067c (MD5) / Made available in DSpace on 2017-03-07T18:32:31Z (GMT). No. of bitstreams: 1 souza_bcc_me_sjrp.pdf: 1030573 bytes, checksum: e3dd1e43953565236359b6d10831067c (MD5) Previous issue date: 2017-02-23 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / O objetivo principal deste trabalho é apresentar detalhadamente um estudo sobre um critério, que aparece na referência Coincidence for maps of spaces with finite group action de D. L. Gonçalves, J. Jaworowski, P. L. Q. Pergher e A. Volovikov, para a existência de (H,G)-coincidências de aplicações cujo contradomínio é um CW-complexo finito Y de dimensão k e cujo domínio é um espaço X paracompacto, Hausdorff, conexo e localmente conexo por caminhos e munido de uma G-ação livre, de modo que exista um inteiro m tal que os grupos i-dimensionais de homologia de X sejam triviais nas dimensões 0<i<m e a cohomologia (m+1)-dimensional de G não seja trivial. Para a realização deste estudo foram necessários alguns resultados da teoria de cohomologia de grupos finitos, com ênfase em grupos de cohomologia periódica segundo a teoria de cohomologia de Tate, alguns resultados da teoria de fibrados e algumas noções da teoria de sequências espectrais cohomológicas. / The mais objective of this work is to present in detail a study about a criterion, which appears in the reference Coincidence for maps of spaces with finite group actions by D. L. Gonçalves, J. Jaworowski, P. L. Q. Pergher and A. Volovikov, for existence of (H,G)-coincidences of maps into a finite CW-complex Y with dimension k and whose domain is a paracompact, Hausdorff, connected and locally pathconnected space X with a free action of G, in a way that there exists an integer m such that the ith-homology group of X is trivial for 0<i<m and the (m+1)th-cohomology group of G is nontrivial. For the study of this criterion were needed some results of the theory of cohomology of finite groups, with emphasis on groups with periodic cohomology according to Tate cohomology theory, some results of the theory of fiber bundles and some notions of the theory of cohomological spectral sequences.

Submitted by Franciele Moreira (francielemoreyra@gmail.com) on 2018-04-17T15:48:39Z No. of bitstreams: 2 Tese - Joelma de Moura Ferreira - 2018.pdf: 58164875 bytes, checksum: c19d300de77be476834ac9c2e7ca8b0e (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2018-04-18T11:17:22Z (GMT) No. of bitstreams: 2 Tese - Joelma de Moura Ferreira - 2018.pdf: 58164875 bytes, checksum: c19d300de77be476834ac9c2e7ca8b0e (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-04-18T11:17:22Z (GMT). No. of bitstreams: 2 Tese - Joelma de Moura Ferreira - 2018.pdf: 58164875 bytes, checksum: c19d300de77be476834ac9c2e7ca8b0e (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-03-21 / Edge bundling is a technique to group, align, coordinate and position the depiction of edges in a graph drawing, so that sets of edges appear to be brought together into shared visual structures, i.e. bundles. The ultimate goal is to reduce clutter to improve how it conveys information. This thesis provides a general formulation for the explicity edge bundling problems, as a formal combinatorial optimization problem. This allows for the definition and comparison of edge bundling problems. In addition, we present four explicity edge bundling optimization problems that address minimizing the total number of bundles, in conjunction with other aspects, as the main goal. An evolutionary edge bundling algorithm is described. The algorithm was successfully tested by solving three related problems applied to real-world instances. The reported experimental results demonstrate the effectiveness and the applicability of the proposed evolutionary algorithm to help resolve edge bundling problems formally defined as optimization models. / A união de arestas em feixes é uma técnica para agrupar, alinhar, coordenar e posicionar a representação de arestas em um desenho de grafo, de modo que os conjuntos de arestas pareçam ser reunidos em estruturas visuais compartilhadas, ou seja, feixes. O objetivo final é reduzir a poluição visual do desenho melhorando a forma como ele transmite informações. Esta tese apresenta uma formulação geral para problemas de união explícita de arestas, como um problema formal de otimização. Essa formulação pode ser usada para definir e comparar problemas de união de arestas. Ainda, são definidos quatro problemas de otimização de união explícita de arestas, que têm por objetivo minimizar o número total de feixes, em conjunto com outros aspectos. Um algoritmo evolucionário é descrito. O algoritmo foi testado com sucesso em três dos problemas relacionados aplicados a instâncias do mundo real. Os resultados experimentais demonstram a eficácia e a aplicabilidade do algoritmo evolutivo proposto para ajudar a resolver problemas de união de arestas em feixes formalmente definidos como um modelo de otimização.

No description available.