The principal goal of this thesis is to develop a systematic method for the design of composite heat sinks (CHSs) that serve as passive and transient cooling devices for microelectronics. This is accomplished by posing the CHS design problem as a topology optimization problem wherein a phase-change material and a high-conductivity material are to be optimally distributed. Two different types of formulations are proposed. The first one aims to maximize the time of operation before a tolerable temperature is reached at the interface between a heat source and the CHS. The second one aims to minimize the maximum temperature across the heating interface for a given time of operation. The two materials are interpolated in topology optimization using the usual mixture law with penalty. The phase-change is modeled using the apparent heat capacity method in which the specific heat is taken as a nonlinear function of the temperature so that the latent heat absorption is accounted for at the melting point. The ensuing new transient topology optimization problem involving an interpolated material property that depends on the state variable is solved using continuous optimization algorithm. The validity of the phase-change modeling is verified with a one dimensional model as well as experimentation. Analytical sensitivity analysis is derived and verified with the finite difference derivatives. Several examples are solved to illustrate the intricacies of the problem and the effectiveness and the limitations of the proposed design method. Prototypes of an intuitively conceived CHS and optimized one are made. An experimental setup is devised to test the two prototypes. Based on the insight gained from the experiments, an improved conduction model is studied to also incorporate convective heat transfer also into the model.

Le génome est très organisé au sein du noyau cellulaire. Cette organisation et plus spécifiquement la localisation et la dynamique des gènes et chromosomes contribuent à l'expression génétique et la différenciation des cellules que ce soit dans le cas de pathologies ou non. L'exploration de cette organisation pourrait dans le futur aider à diagnostiquer et identifier de nouvelles cibles thérapeutiques. La conformation des chromosomes peut être analysée grâce au marquage ADN sur plusieurs sites et aux mesures de distances entre ces différents marquages fluorescents. Dans ce contexte, l'organisation spatiale du chromosome III de levure a montré que les deux types de cellules, MATa et MATalpha, sont différents. Par contre, les données issues de l'imagerie electronique sont bruitées à cause de la résolution des systèmes de microscope et du fait du caractère vivant des cellules observées. Dans cette thèse, nous nous intéressons au développement de méthodes de classification pour différencier les types de cellules sur la base de mesures de distances entre 3 loci du chromosome III et d'une estimation du bruit. Dans un premier temps, nous nous intéressons de façon générale aux problèmes de classification binaire à l'aide de SVM de grandes tailles et passons en revue les algorithmes d'optimisation stochastiques du premier ordre. Afin de prendre en compte les incertudes, nous proposons un modèle d'apprentissage qui ajuste sa robustesse en fonction du bruit. La méthode évite les situations où le modèle est trop conservatif et que l'on rencontre parfois avec les formulations SVM robustes. L'amplitude des pertubations liées au bruit qui sont incorporées dans le modèle est controllée par l'optimisation d'une erreur de généralisation. Aucune hypothèse n'est faite sur la distribution de probabilité du bruit. Seule une borne estimée des pertubations est nécessaire. Le problème peut s'écrire sous la forme d'un programme biniveaux de grande taille. Afin de le résoudre, nous proposons un algorithme biniveau qui réalise des déplacements stochastiques très peu coûteux et donc adapté aux problèmes de grandes tailles. La convergence de l'algorithme est prouvée pour une classe générale de problèmes. Nous présentons des résultats numériques très encourageants qui confirment que la technique est meilleure que l'approche SOCP (Second Order Cone Programming) pour plusieurs bases de données publiques. Les expériences numériques montrent également que la nonlinéarité additionnelle générée par l'incertitude sur les données pénalise la classification des chromosomes et motivent des recherches futures sur une version nonlinéaire de la technique proposée. Enfin, nous présentons également des résultats numériques de l'algorithme biniveau stochastique pour la sélection automatique de l'hyperparamètre de pénalité dans les SVM. L'approche évite les coûteux calculs que l'on doit inévitablement réaliser lorsque l'on effectue une validation croisée sur des problèmes de grandes tailles. / The genome is highly organized within the cell nucleus. This organization, in particular the localization and dynamics of genes and chromosomes, is known to contribute to gene expression and cell differentiation in normal and pathological contexts. The exploration of this organization may help to diagnose disease and to identify new therapeutic targets. Conformation of chromosomes can be analyzed by distance measurements of distinct fluorescently labeled DNA sites. In this context, the spatial organization of yeast chromosome III was shown to differ between two cell types, MATa and MATa. However, imaging data are subject to noise, due to microscope resolution and the living state of yeast cells. In this thesis, the aim is to develop new classification methods to discriminate two mating types of yeast cells based on distance measurements between three loci on chromosome III aided by estimation the bound of the perturbations. We first address the issue of solving large scale SVM binary classification problems and review state of the art first order optimization stochastic algorithms. To deal with uncertainty, we propose a learning model that adjusts its robustness to noise. The method avoids over conservative situations that can be encountered with worst case robust support vector machine formulations. The magnitude of the noise perturbations that is incorporated in the model is controlled by optimizing a generalization error. No assumption on the distribution of noise is taken. Only rough estimates of perturbations bounds are required. The resulting problem is a large scale bi-level program. To solve it, we propose a bi-level algorithm that performs very cheap stochastic gradient moves and is therefore well suited to large datasets. The convergence is proven for a class of general problems. We present encouraging experimental results confirming that the technique outperforms robust second order cone programming formulations on public datasets. The experiments also show that the extra nonlinearity generated by the uncertainty in the data penalizes the classification of chromosome data and advocates for further research on nonlinear robust models. Additionally, we provide the experimenting results of the bilevel stochastic algorithm used to perform automatic selection of the penalty parameter in linear and non-linear support vector machines. This approach avoids expensive computations that usually arise in k-fold cross validation.

Atualmente as estruturas estão sendo avaliadas para um maior número de ações em relação há algumas décadas. Esta melhoria ao longo da fase de concepção é dada devido ao fato de que está se tornando mais competitivo o fornecimento de estruturas leves e esbeltas, sendo solicitados, cada vez mais, projetos com menor custo de implantação. Devido a isto, é necessário avaliar as estruturas não apenas sujeitas a cargas estáticas, mas também a carregamentos dinâmicos. As ações dinâmicas que atuam sobre uma estrutura podem ser muito mais prejudiciais do que as estáticas quando não são bem consideradas e dimensionadas. Ações dinâmicas podem ser provenientes de tremores de terra, vento, equipamentos em funcionamento, deslocamento de pessoas, veículos em movimento, motores desbalanceados, entre outras fontes, o que pode causar vibrações na estrutura, podendo levar a mesma ao colapso. A fim de controlar e reduzir as amplitudes de vibração, entre outras alternativas é possível a instalação de amortecedores de massa sintonizado (AMS), que é um dispositivo de controle passivo. O AMS tem várias vantagens, tais como a grande capacidade de reduzir a amplitude de vibração, fácil instalação, baixa manutenção, baixo custo, entre outras. Para se obter a melhor relação custo-benefício, ou seja, a maior redução de amplitude aliada a um menor número de amortecedores ou a uma menor massa, a otimização dos parâmetros do AMS tornase fundamental. Neste contexto, este trabalho visa, através de simulação numérica, propor um método para otimizar parâmetros de AMSs quando estes devem ser instalados em edifícios submetidos à excitação sísmica. Inicialmente é considerado apenas um único AMS instalado no topo do edifício e em seguida também são feitas simulações com múltiplos AMSs (MAMS), e por fim são descartados os AMSs desnecessários, obtendo assim a melhor resposta da estrutura. Para tanto, uma rotina computacional é desenvolvida em MatLab usando o método de integração direta das equações de movimento de Newmark para determinar a resposta dinâmica da estrutura. Para fins de análise podem ser considerados tanto sismos reais quanto artificiais. Os acelerogramas artificias são gerados a partir do espectro proposto por Kanai e Tajimi. Primeiramente, a estrutura é analisada somente com o seu amortecimento próprio para fins comparativos e de referência. Em seguida, a otimização do ou dos AMSs é feita, na qual a função objetivo é minimizar o deslocamento máximo no topo do edifício, e as variáveis de projeto, são a relação de massas (AMS - Estrutura), rigidez e amortecimento do ou dos AMSs. Para a otimização são utilizados os algoritmos Firefly Algotithm e Backtracking Search Optimization Algorithm. De acordo com as configurações do AMS, após a otimização dos seus parâmetros são determinadas as novas respostas dinâmicas da estrutura. Finalmente, pode-se observar que o método proposto foi capaz de otimizar os parâmetros do ou dos AMSs, reduzindo consideravelmente as respostas da estrutura após a instalação do mesmo, minimizando o risco de dano e colapso do edifício. Desta forma, este trabalho mostra que é possível projetar AMS e MAMS de forma econômica e eficaz. / Currently, structures are being evaluated for a greater number of actions when compared to a few decades ago. This improvement in designing stage is happening because projects providing lightweight and slender structures, with lower implantation costs, are being more requested. Thus, evaluating structures not only subjected to static loads, but also to dynamic loads has become necessary. Dynamic loads acting on a structure are more damaging than static loads, if they are not well considered and dimensioned. Dynamic loads could occur from earthquakes, wind, equipment, movement of people or vehicles, among other sources, which cause vibrations in structures and may lead to a collapse. Tuned mass damper (TMD), a passive control device, can be installed as an alternative to reduce vibration amplitudes. TMD has several advantages, such as large capacity to reduce amplitude of vibration, easy installation, low maintenance, low cost, among others. Optimizing TMD parameters is fundamental for obtaining best cost-benefit relation, i.e., greater amplitude reduction along with lower number of dampers or lower mass. In this context, this study aims at proposing, through numerical simulation, a method for optimizing TMD parameters when installing them on buildings under seismic excitation. Initially, a single-TMD case is considered, then simulations with multiple-TMDs (MTMDs) are run; lastly, unnecessary TMDs are discarded, obtaining the best structural response. For this purpose, a computational routine is developed on MatLab using Newmark direct integration method for equations of motion to determine the dynamic structural response. Both real and artificial earthquakes are considered for purposes of analysis. Artificial accelerograms are generated from proposed Kanai-Tajimi spectrum. First, structure is analyzed only with its own damping for comparison and reference. Second, a single or multiple-TMD optimization is carried out, in which the objective function is to minimize the maximum displacement at the top of the building, and the design variables are modal mass ratio (Structure-TMD), stiffness and damping of a single or multiple-TMD. Firefly and Backtracking Optimization algorithms are used for optimization. According to TMD settings, new dynamic structural responses are determined after optimizing parameters. Finally, the proposed method could optimize parameters of single or multiple-TMDs, considerably reducing structural responses after their installation, minimizing the risk of damage and building collapse. Thus, this study shows the possibility of designing TMDs or MTMDs both economically and effectively.

Les recherches de cette thèse portent sur la mise en oeuvre des méthodes de la rechercheopérationnelle (RO) pour la conception de circuits numériques dans le domaine du traitementdu signal et de l’image, plus spécifiquement pour des applications multimédia et de visionembarquée.Face à la problématique de “Memory Wall”, les concepteurs de systèmes de vision embarquée,Mancini et al. (Proc.DATE, 2012), ont proposé un générateur de hiérarchies mémoiresad-hoc dénommé Memory Management Optimization (MMOpt). Cet atelier de conception estdestiné aux traitements non-linéaires afin d’optimiser la gestion des accès mémoires de cestraitements. Dans le cadre de l’outil MMOpt, nous abordons la problématique d’optimisationliée au fonctionnement efficace des circuits de traitement d’image générés par MMOpt visantl’amélioration des enjeux de performance (contrainte temps-réel), de consommation d’énergieet de coût de production (contrainte d’encombrement).Ce problème électronique a été modélisé comme un problème d’ordonnancement multiobjectif,appelé 3-objective Process Scheduling and Data Prefetching Problem (3-PSDPP), reflétantles 3 principaux enjeux électroniques considérés. À notre connaissance, ce problème n’apas été étudié avant dans la littérature de RO. Une revue de l’état de l’art sur les principaux travauxliés à cette thèse, y compris les travaux antérieurs proposés par Mancini et al. (Proc.DATE,2012) ainsi qu’un bref aperçu sur des problèmes voisins trouvés dans la littérature de RO,a ensuite été faite. En outre, la complexité de certaines variantes mono-objectif du problèmed’origine 3-PSDPP a été établie. Des approches de résolution, y compris les méthodes exactes(PLNE) et les heuristiques constructives, sont alors proposées. Enfin, la performance de cesméthodes a été comparée par rapport à l’algorithme actuellement utilisé dans l’outil MMOpt,sur des benchmarks disponibles dans la littérature ainsi que ceux fournis par Mancini et al.(Proc.DATE, 2012).Les solutions obtenues sont de très bonne qualité et présentent une piste prometteuse pouroptimiser les performances des hiérarchies mémoires produites par MMOpt. En revanche, vuque les besoins de l’utilisateur de l’outil sont contradictoires, il est impossible de parler d’unesolution unique en optimisant simultanément les trois critères considérés. Un ensemble debonnes solutions de compromis entre ces trois critères a été fourni. L’utilisateur de l’outilMMOpt peut alors décider de la solution qui lui est la mieux adaptée. / The research of this thesis focuses on the application of the Operations Research (OR)methodology to design new optimization algorithms to enable low cost and efficient embeddedvision systems, or more generally devices for multimedia applications such as signal and imageprocessing.The design of embedded vision systems faces the “Memory Wall” challenge regarding thehigh latency of memories holding big image data. For the case of non-linear image accesses, onesolution has been proposed by Mancini et al. (Proc. DATE 2012) in the form of a software tool,called Memory Management Optimization (MMOpt), that creates an ad-hoc memory hierarchiesfor such a treatment. It creates a circuit called a Tile Processing Unit (TPU) that containsthe circuit for the treatment. In this context, we address the optimization challenge set by theefficient operation of the circuits produced by MMOpt to enhance the 3 main electronic designcharacteristics. They correspond to the energy consumption, performance and size/productioncost of the circuit.This electronic problem is formalized as a 3-objective scheduling problem, which is called3-objective Process Scheduling and Data Prefetching Problem (3-PSDPP), reflecting the 3 mainelectronic design characteristics under consideration. To the best of our knowledge, this problemhas not been studied before in the OR literature. A review of the state of the art, including theprevious work proposed by Mancini et al. (Proc.DATE, 2012) as well as a brief overview onrelated problems found in the OR literature, is then made. In addition, the complexity of someof the mono-objective sub-problems of 3-PSDPP problem is established. Several resolutionapproaches, including exact methods (ILP) and polynomial constructive heuristics, are thenproposed. Finally, the performance of these methods is compared, on benchmarks available inthe literature, as well as those provided by Mancini et al. (Proc.DATE, 2012), against the onecurrently in use in the MMOpt tool.The results show that our algorithms perform well in terms of computational efficiency andsolution quality. They present a promising track to optimize the performance of the TPUs producedby MMOpt. However, since the user’s needs of the MMOpt tool are contradictory, such aslow cost, low energy and high performance, it is difficult to find a unique and optimal solutionto optimize simultaneously the three criteria under consideration. A set of good compromisesolutions between these three criteria was provided. The MMOpt’s user can then choose thebest compromise solution he wants or needs.

Atualmente as estruturas estão sendo avaliadas para um maior número de ações em relação há algumas décadas. Esta melhoria ao longo da fase de concepção é dada devido ao fato de que está se tornando mais competitivo o fornecimento de estruturas leves e esbeltas, sendo solicitados, cada vez mais, projetos com menor custo de implantação. Devido a isto, é necessário avaliar as estruturas não apenas sujeitas a cargas estáticas, mas também a carregamentos dinâmicos. As ações dinâmicas que atuam sobre uma estrutura podem ser muito mais prejudiciais do que as estáticas quando não são bem consideradas e dimensionadas. Ações dinâmicas podem ser provenientes de tremores de terra, vento, equipamentos em funcionamento, deslocamento de pessoas, veículos em movimento, motores desbalanceados, entre outras fontes, o que pode causar vibrações na estrutura, podendo levar a mesma ao colapso. A fim de controlar e reduzir as amplitudes de vibração, entre outras alternativas é possível a instalação de amortecedores de massa sintonizado (AMS), que é um dispositivo de controle passivo. O AMS tem várias vantagens, tais como a grande capacidade de reduzir a amplitude de vibração, fácil instalação, baixa manutenção, baixo custo, entre outras. Para se obter a melhor relação custo-benefício, ou seja, a maior redução de amplitude aliada a um menor número de amortecedores ou a uma menor massa, a otimização dos parâmetros do AMS tornase fundamental. Neste contexto, este trabalho visa, através de simulação numérica, propor um método para otimizar parâmetros de AMSs quando estes devem ser instalados em edifícios submetidos à excitação sísmica. Inicialmente é considerado apenas um único AMS instalado no topo do edifício e em seguida também são feitas simulações com múltiplos AMSs (MAMS), e por fim são descartados os AMSs desnecessários, obtendo assim a melhor resposta da estrutura. Para tanto, uma rotina computacional é desenvolvida em MatLab usando o método de integração direta das equações de movimento de Newmark para determinar a resposta dinâmica da estrutura. Para fins de análise podem ser considerados tanto sismos reais quanto artificiais. Os acelerogramas artificias são gerados a partir do espectro proposto por Kanai e Tajimi. Primeiramente, a estrutura é analisada somente com o seu amortecimento próprio para fins comparativos e de referência. Em seguida, a otimização do ou dos AMSs é feita, na qual a função objetivo é minimizar o deslocamento máximo no topo do edifício, e as variáveis de projeto, são a relação de massas (AMS - Estrutura), rigidez e amortecimento do ou dos AMSs. Para a otimização são utilizados os algoritmos Firefly Algotithm e Backtracking Search Optimization Algorithm. De acordo com as configurações do AMS, após a otimização dos seus parâmetros são determinadas as novas respostas dinâmicas da estrutura. Finalmente, pode-se observar que o método proposto foi capaz de otimizar os parâmetros do ou dos AMSs, reduzindo consideravelmente as respostas da estrutura após a instalação do mesmo, minimizando o risco de dano e colapso do edifício. Desta forma, este trabalho mostra que é possível projetar AMS e MAMS de forma econômica e eficaz. / Currently, structures are being evaluated for a greater number of actions when compared to a few decades ago. This improvement in designing stage is happening because projects providing lightweight and slender structures, with lower implantation costs, are being more requested. Thus, evaluating structures not only subjected to static loads, but also to dynamic loads has become necessary. Dynamic loads acting on a structure are more damaging than static loads, if they are not well considered and dimensioned. Dynamic loads could occur from earthquakes, wind, equipment, movement of people or vehicles, among other sources, which cause vibrations in structures and may lead to a collapse. Tuned mass damper (TMD), a passive control device, can be installed as an alternative to reduce vibration amplitudes. TMD has several advantages, such as large capacity to reduce amplitude of vibration, easy installation, low maintenance, low cost, among others. Optimizing TMD parameters is fundamental for obtaining best cost-benefit relation, i.e., greater amplitude reduction along with lower number of dampers or lower mass. In this context, this study aims at proposing, through numerical simulation, a method for optimizing TMD parameters when installing them on buildings under seismic excitation. Initially, a single-TMD case is considered, then simulations with multiple-TMDs (MTMDs) are run; lastly, unnecessary TMDs are discarded, obtaining the best structural response. For this purpose, a computational routine is developed on MatLab using Newmark direct integration method for equations of motion to determine the dynamic structural response. Both real and artificial earthquakes are considered for purposes of analysis. Artificial accelerograms are generated from proposed Kanai-Tajimi spectrum. First, structure is analyzed only with its own damping for comparison and reference. Second, a single or multiple-TMD optimization is carried out, in which the objective function is to minimize the maximum displacement at the top of the building, and the design variables are modal mass ratio (Structure-TMD), stiffness and damping of a single or multiple-TMD. Firefly and Backtracking Optimization algorithms are used for optimization. According to TMD settings, new dynamic structural responses are determined after optimizing parameters. Finally, the proposed method could optimize parameters of single or multiple-TMDs, considerably reducing structural responses after their installation, minimizing the risk of damage and building collapse. Thus, this study shows the possibility of designing TMDs or MTMDs both economically and effectively.

A tarefa de planejamento de trajetórias de robôs móveis autônomos consiste em determinar objetivos intermediários para que um robô seja capaz de partir de sua localização inicial e alcançar seu objetivo final. Além do planejamento, é importante definir um método de controle da navegação (seguimento da trajetória) do robô para que ele seja capaz de realizar seu trajeto de forma segura. Este projeto propõe uma abordagem híbrida para planejamento exploratório e execução de trajetórias de robôs móveis autônomos em ambientes indoor. Para o planejamento de trajetória, foram investigados algoritmos de busca em espaço de estados, dando ênfase ao uso de algoritmos evolutivos e algoritmos de otimização por colônia de formigas para a descoberta e otimização da trajetória. O controle da navegação é realizado por meio de comportamentos locais reativos, baseado na exploração e uso de mapas topológicos, os quais permitem uma maior flexibilidade em termos de definição da localização da posição do robô móvel e sobre os detalhes do mapa do ambiente (mapas com informações aproximadas e não métricos). Assim, foi proposto e desenvolvido um método robusto capaz de planejar, mapear e explorar um caminho ótimo ou quase ótimo para que o robô possa navegar e alcançar seu objetivo de forma segura, com pouca informação prévia do ambiente ou mesmo sobre sua localização. Além disso, o robô pode reagir a ambientes com alterações dinâmicas em sua estrutura, considerando por exemplo, elementos dinâmicos como portas que possam ser abertas ou fechadas e passagens que são obstruídas. Por fim, foram realizados diversos testes e simulações a fim de validar o método proposto, com a avaliação da qualidade das soluções encontradas e comparação com outras abordagens tradicionais de planejamento de trajetórias (algoritmos A* e D*). / The task of planning path for autonomous mobile robots consists in determine intermediary goals in order to allow a robot be able to leave its initial location and reach its final goal. Besides the planning, it is important to define a method of navigation control (the trajectory following) of the robot for it be able to do its path safely. This project proposes a hybrid approach to path planning and execution of an autonomous mobile robot in indoor environments. For the path planning, search algorithms in state space have been investigated, with emphasis in evolutionary algorithms and ant colony optimization algorithms for the trajectory search and optimization. The navigation control is done by local reactive behaviors, based on topological maps, which allow more flexibility concerning localization definition of position of the mobile robot and about the details of the environment map (maps with approximate information and not metric). Thus, a robust method able to plan an optimum or almost optimum path for the robot to reach its goal safely has been proposed, with little previous information of the environment. Furthermore, the robot can react to dynamic elements in the environment structure, concerning, for example, dynamic elements such as doors that can be opened or closed and ways that are blocked. Finally, several tests and simulations has been carried out to validate the proposed method, with evaluation of the solutions quality and comparison with others traditional approaches for the path planning task (A* and D* algorithms).

O crescimento do tráfego de serviços de telecomunicações tem aumentado o consumo de energia e, em consequência, aumentado as emissões de CO2 que tem efeitos nocivos sobre o meio ambiente. É assim que a economia de energia torna-se um fator chave no planejamento de redes de telecomunicações. Para garantir a disponibilidade e confiabilidade, as redes possuem arquitetura redundante e são projetadas para suportar a demanda de pico de tráfego. Redes com mecanismos de proteção como proteção dedicada de caminhos (DPP), proveem caminhos alternativos para cada demanda de conexão. Os elementos da rede que suportam esses caminhos estão em estado ativo (consumindo energia), apesar de, na maior parte do tempo, não transportarem tráfego efetivo. Um método para diminuir o gasto de energia é utilizar roteamento adaptado à carga real de tráfego baseado em modo suspenso (estado de baixo consumo de energia que pode passar a estado ativo rapidamente). Assim, o tráfego é roteado com vistas à maximizar a quantidade de componentes que são parte de caminhos de proteção, que podem ser postos em modo suspenso. Neste trabalho, as redes usadas para os testes são a rede europeia Cost239, a rede estadunidense UsNet e a rede brasileira Ipê. Abordamos o problema de economia de energia em redes WDM com DPP através de quatro estratégias de roteamento. Cada uma tem objetivos diferentes, a Shortest Path-DPP (SP-DPP) faz o roteamento por caminho mais curto, a Energy Aware-DPP (EA-DPP) aloca as demandas por enlaces que estejam ativos, a Energy Aware-DPP with Mixing (EA-DPP-MixS) evita que caminhos principais sejam roteados por enlaces que já são parte de caminhos de proteção e a Energy Aware-DPP with Differentation (EA-DPP-Dif) evita a mistura de caminhos por um mesmo enlace. Em nossas simulações computacionais observamos que a EA-DPP-Dif economiza energia de maneira eficiente, mas a probabilidade de bloqueio aumenta. A EA-DPP-MixS diminui o bloqueio em detrimento da energia economizada. Já a SP-DPP e a EA-DPP são menos eficientes na diminuição da energia consumida. É assim que propomos um roteamento com busca de recursos mais ampla, usando cada uma das estratégias. A proposta será chamada de roteamento intensivo. A EA-DPP-Dif-Intensivo diminui a probabilidade de bloqueio e economiza energia mediante modo suspenso. Neste trabalho, analisamos o desempenho das estratégias para cada uma das redes e avaliamos o impacto da energia economizada sobre a probabilidade de bloqueio. A proposta de roteamento i>intensivo diminui a energia consumida em até 50%, diminuindo a probabilidade de bloqueio. Porém, os resultados estão diretamente relacionados com a carga de rede e as características particulares da topologia de cada rede. / The growth of data traffic in telecommunication networks has increased energy consumption and hence increased CO2 emissions, with harmful effects on the environment. Thus, energy saving becomes a key and a differential factor when planning telecommunication networks. In order to guarantee availability and reliability, core networks have redundant architecture and are designed to support peak-hour traffic demand. Networks with dedicated path protection (DPP) mechanisms provide alternative paths for each connection request. Network elements supporting these paths are in active state (consuming energy), although most of the time they dont carry traffic. One technique to decrease energy waste is by adaptive real traffic routing using sleep mode (a low energy consumption state which is able to rapidly change to an active state). Thus, traffic is routed in order to maximize the amount of network components used by protection paths, which can be set in sleep mode. In this work, European Cost239, American UsNet and Brazilian Ipê networks were used in computational simulations. We addressed the energy saving problem in WDM networks with DPP through four routing strategies, each with different goals. The Shorthest Path-Dedicated Path Protection (SP-DPP) technique uses shortest path for routing, Energy Aware-Dedicated Path Protection (EA-DPP) allocates demands in active links, Energy Aware-Dedicated Path Protection with Mixing (EA-DPP-MixS) prevents primary paths to be formed by links that are already part of the protection paths and Energy Aware-Dedicated Path Protection with Differentation (EA-DPP-Dif) prevents mixing primary and protection paths through the same link. We observe that EA-DPP-Dif efficiently saved energy, however blocking probability has increased. EA-DPP-MixS reduced blocking rather than saved energy. At least, SP-DPP and EA-DPP are less efficient in reducing energy consumption. Hence, we propose a wider resource search routing, the in-depth routing, using each of these strategies. Thus, EA-DPP-Dif-In-depth decreased blocking probability while maintaining energy saving through sleep mode. In this work, we analyze the strategies performance for each network and evaluate the impact of energy saved on the blocking probability. Our in-depth routing strategy reduced the energy consumption up to 50%, decreasing blocking probability. However, the results are directly related with the network load and the specific properties of each network topology.

Nos ultimos anos o uso de tecnicas de aprendizado computacional tornou se uma das tarefas comumente realizadas, pois tem inumeras aplicacoes de reconhecimento de padroes, tais como: reco- nhecimento de voz, classificacao de texto, reconhecimento facial, diagnostico por imagens medicas, entre outras. Dessa forma, um grande numero de tecnicas que lidam com este tipo de problema tem sido desenvolvido ate o momento. Neste trabalho apresentamos uma alternativa para melhorar a taxa acerto de classificacao do classificador binario SLS, que apresentou resultados comparaveis com as SVMs. Nesse metodo, o Gradiente Descendente e utilizado para otimizar a posicao final dos conjuntos de segmentos de reta que representarao cada classe. Embora convirja rapidamente a um valor otimo, muitas vezes e possivel o algoritmo parar em uma regiao de otimos locais, que nao representa o minimo global. Dado esse problema, foram utilizados diferentes algoritmos evolutivos em combinacao com o Gradiente Descendente a fim de melhorar a acuracia do classificador SLS. Adicionalmente a aplicacao de algoritmos evolutivos na fase de treinamento do classificador SLS, foram exploradas duas propostas: (i) explorar o uso de diferente numero de segmentos de reta para representar a distribuicao de dados de cada classe. Dado que no algoritmo original do metodo SLS o numero de segmentos de reta e igual para cada classe, o qual pode significar alguma perda de acuracia ou sobreposicao dos segmentos de reta; (ii) estimar a melhor combinacao de segmentos de reta a serem usados para cada classe. O uso de diferentes quantidades de segmentos de reta por classe pode ser de ajuda na obtencao de melhores porcentagens de acerto, mas determinar uma quantidade otima que permita representar cada classe, e um trabalho dificil. Assim, usamos o algoritmo X-Means, que e um algoritmo de agrupamento, para estimar o numero de segmentos de reta. As propostas exibiram bons resultados que possibilitam a aplicacao do classificador SLS, com um algoritmo de treinamento hibrido, em problemas reais. / During the past years, the use of machine learning techniques have become into one of the most frequently performed tasks, due to the large amount of pattern recognition applications such as: voice recognition, text classification, face recognition, medical image diagnosis, among others. Thus, a great number of techniques dealing with this kind of problem have been developed until now. In this work, we propose an alternative training algorithm to improve the accuracy of the SLS binary Classifier, which produces good results that can be compared to Support Vector Machines. In that classifier, the Gradient Descent method has been used to optimize the final positions of two sets of straight line segments that represent each class. Although, this method quickly converges to an optimum, it is possible that the algorithm stops at a local optimum region, which does not guarantee a global minimum. Given that problem, we combine evolutive optimization algorithms with the gradient descent method to improve the accuracy of the SLS Classifier. In addition to our proposal of using evolutive algorithms, we also developed two proposals: (i) we explore the use of different number of straight line segments to represent the data distribution. Since the original SLS classifier algorithm uses the same number of segments for each class, which could lead to a loss of accuracy or straight line segments overlapping. So, using different number of segments could be the way to improve the accuracy; (ii) estimate the best combination of straight line segments to represent each class. Finding an optimal combination, can be a very difficult problem, so we propose the X-Means algorithm to determine the number of segments. The proposed methodology showed good results which can be used to solve some other real problems with the SLS classifier using the proposed hybrid training algorithm.

<pre>Coherent lidar uses a chirped laser pulse for 3D imaging of distant targets.However, existing coherent lidar image reconstruction methods do not account for the system's aperture, resulting in sub-optimal resolution.Moreover, these methods use majorization-minimization for computational efficiency, but do so without a theoretical treatment of convergence.<br> <br>In this work, we present Coherent Lidar Aperture Modeled Plug-and-Play (CLAMP) for multi-look coherent lidar image reconstruction.CLAMP uses multi-agent consensus equilibrium (a form of PnP) to combine a neural network denoiser with an accurate physics-based forward model.CLAMP introduces an FFT-based method to account for the effects of the aperture and uses majorization of the forward model for computational efficiency.We also formalize the use of majorization-minimization in consensus optimization problems and prove convergence to the exact consensus equilibrium solution.Finally, we apply CLAMP to synthetic and measured data to demonstrate its effectiveness in producing high-resolution, speckle-free, 3D imagery.</pre><p></p>

Η πρόσφατη τεχνολογική ανάπτυξη των οπτικών ενισχυτών, πολυπλεκτών/αποπλεκτών, οπτικών διακοπτών καθώς και άλλων οπτικών συσκευών μας οδηγεί στο να ελπίζουμε ότι σύντομα στο μέλλον θα υλοποιηθεί ένα πλήρες οπτικό (all optical), WDM (wavelength division multiplexing) δίκτυο που να ικανοποιεί και την ανάγκη για μεγάλα μεγέθη χωρητικότητας. Σε ένα τέτοιο δίκτυο η μετατροπή του οπτικού σήματος σε ηλεκτρονικό και εκ νέου στο οπτικό (ΟΕΟ) δεν θα χρησιμοποιείται στους ενδιάμεσους κόμβους, και αυτό συμβάλει σε οικονομικότερες υλοποιήσεις των οπτικών δικτύων. Σε ένα WDM δρομολογούμενο δίκτυο, τα δεδομένα μεταφέρονται μέσω ενός οπτικού καναλιού, lightpath, στους κόμβους του δικτύου που συνδέονται με οπτικές ίνες. Στις πλείστες των περιπτώσεων, κατά την άφιξη ενός lightpath σε κάποιο κόμβο, εφαρμόζεται σε αυτό οπτικό-ηλεκτρονική μετατροπή και αντίστροφα, ούτως ώστε το σήμα να αναδημιουργηθεί λόγω των απωλειών που υπέστη κατά την μεταφορά, ή ακόμη για να αναλυθεί από ενδιάμεσες ηλεκτρονικές συσκευές. Στα μη πλήρη οπτικά δίκτυα, η μεταφορά των δεδομένων γίνεται από κόμβο σε κόμβο κατά μήκος του δικτύου, ούτως ώστε το οπτικό σήμα να ενισχύεται και να αναγεννάτε μέσω της OEO επεξεργασίας. Παρ’ όλα αυτά, η κάθε ενδιάμεση ανάλυση του θέματος σε ένα τέτοιο δίκτυο προϋποθέτει πολύ μεγάλα κόστη λόγω των πολλών συσκευών που απαιτούνται για τη OEO επεξεργασία. Το γεγονός αυτό μας οδηγεί στα ημί-πλήρη δίκτυα όπου η ενίσχυση και αναγέννηση του θέματος δε γίνεται σε όλους τους ενδιάμεσους κόμβους αλλά σε μερικούς από αυτούς. Ο τελικός στόχος όμως είναι η απαλοιφή της ηλεκτρονικής μετατροπής και αυτό οδηγεί στην υλοποίηση των πλήρως οπτικών δικτύων. Στα πλήρη οπτικά δίκτυα, ένα σήμα που μεταδίδεται παραμένει, για όλο το lightpath, στο οπτικό επίπεδο. Έτσι, το πλήρες οπτικό δίκτυο μπορεί να απαλείψει την ασύμφορη OEO μετατροπή. Η αναζήτηση των κατάλληλων μονοπατιών με τα κατάλληλα μήκη κύματος που θα ικανοποιούσε ένα πλήρες οπτικό δίκτυο το οποίο δρομολογείται από ligthpaths, ονομάζεται Routing and Wavelength Assignment (RWA) και αποτελεί ένα από τα σημαντικότερα ζητήματα για το σωστό σχεδιασμό των οπτικών δικτύων τέτοιου είδους. Το πρόβλημα γίνεται ιδιαίτερα πολύπλοκο όταν στην τελική απόφαση θα πρέπει να συμπεριληφθούν και τα χαρακτηριστικά του φυσικού επιπέδου του δικτύου, όπως εξασθένιση του σήματος, μη γραμμικά φαινόμενα, διασπορά κ.ά, η συμβολή των οποίων στην τελική δρομολόγηση δεν θεωρείται αμελητέα (Impairment Aware Routing and Wavelength Assignment, ΙΑ-RWA). Σε αυτή την εργασία μελετάται το RWA πρόβλημα και προτείνεται ένας μονού στόχου γενετικός αλγόριθμος (Single Objective Genetic Algorithm - SOGA), ο οποίος επιλύει ικανοποιητικά το πρόβλημα θεωρώντας στατική κίνηση. Επιπλέον τονίζεται η σημασία των φυσικών παραμέτρων του προβλήματος και πως αυτές επηρεάζουν την απόδοση του πλήρους οπτικού δικτυου. Στη συνέχεια προτείνεται ένας νέος, πολλαπλών στόχων γενετικός αλγόριθμος (multi objective genetic algorithm – MOGA) ο οποίος βελτιστοποιεί τις λύσεις του προβλήματος ικανοποιητικά λαμβάνοντας ταυτόχρονα υπόψη, με έμμεσο τρόπο, και τις φυσικές παραμέτρους. Επίσης προτείνεται και ένας μονού στόχου γενετικός αλγόριθμος οποίος χρησιμοποιεί ένα εργαλείο αποτίμηση της ποιότητας μετάδοσης (Q-TOOL) σαν μέτρο κατά τη διαδικασία εύρεσης ικανοποιητικής λύσης. Το υπόλοιπο της εργασίας οργανώνεται ως ακολούθως: Στην ενότητα 2 παρουσιάζεται μια σύντομη αναφορά στα WDM δίκτυα καθώς και η περιγραφή του RWA και IA-RWA προβλήματος, ενώ στην ενότητα 3 παρουσιάζεται η πρόταση επίλυσης του RWA προβληματος με τη χρήση γενετικών αλγορίθμων. Ακολουθεί στην ενότητα 4 η πρότασή μας για επίλυση του IA-RWA προβλήματος με τη χρήση Multi-objective διαδικασιών βελτιστοποίησης, καθώς και η βελτιστοποίηση του προβλήματος με τη χρήση του Q-TOOL. Τέλος στην ενότητα 5 συνοψίζουμε την εργασία και παρουσιάζουμε τα συμπεράσματα. / The recent development of optical amplifiers, multiplexers / de-multiplexers, optical switches and other optical devices leads us to hope that soon in future all optical, WDM (wavelength division multiplexing) networks will be implemented which that will satisfy the needs for large capacity. In such networks a viable conversion of the optical -> Electronic and back to optical (OEO) will not be used at intermediate nodes, and this will contribute to efficient and economical implementation. The search for the appropriate paths with the appropriate wavelengths that meet the requirement in all optical networks is called Routing and Wavelength Assignment (RWA) and is one of the most important issues for proper design of such optical networks. The problem becomes particularly complex when the final decision should include the characteristics of the physical layer of the network, such as attenuation of the signal, nonlinear effects, dispersion, etc., whose contribution to the final result is not considered negligible (Impairment Aware Routing and Wavelength Assignment,IA-RWA). This work studies the RWA problem considering static traffic, and proposes a single-objective genetic algorithm (Single Objective Genetic Algorithm - SOGA), which resolves the problem satisfactorily. Furthermore the work stresses the importance of physical parameters of the problem and how these affect the performance of the all optical networks, and proposes a new, multi-objective genetic algorithm (MOGA) which optimizes the solution of IA-RWA problem adequately taking into account indirectly, and the physical impairments that affect the quality of the signal. In addition, a single objective genetic algorithm is proposed that uses a tool to assess the quality of the transmission signal (Q-TOOL), as a benchmark, in the process of optimization of the solution to the IA-RWA problem.