As estruturas aeronáuticas estão sujeitas a diversas solicitações, devido principalmente às interações com o escoamento aerodinâmico, que podem causar distúrbios e vibrações, comprometendo seu desempenho. Diversas pesquisas vêm sendo realizadas para solucionar estes problemas. Dentre elas está o uso de atuadores e sensores piezelétricos integrados na estrutura, que juntamente com um sistema de controle passa a ser denominada estrutura inteligente, a qual promove o controle ativo de vibrações garantindo um aumento no desempenho. O objetivo deste trabalho é obter parâmetros ótimos de um controlador não convencional baseado na lógica difusa para controle de vibrações em uma viga com atuadores e sensores piezelétricos. A viga e elementos piezelétricos são modelados pelo método de elementos finitos utilizando o princípio variacional eletromecânico. O sistema de controle difuso, o qual está se tornando amplamente utilizado principalmente devido à sua capacidade de representar sistemas não lineares e complexos, é baseado nos modelos difusos de Mamdani e Takagi-Sugeno-Kang. A otimização é feita através de algoritmo genético que é um processo de procura probabilística baseado nas leis de seleção natural influenciadas pelas teorias de Charles Darwin. São otimizados os valores dos ganhos de controle, bem como os suportes dos conjuntos difusos da base de conhecimento. São feitas comparações com o controlador difuso obtido por processo de ajuste manual. / Aeronautical structures are subject to a variety of loads, due mainly to the iteration with the aerodynamic flow that can present disturbances, compromising their performance. Various researches have been carried out to solve these problems. Among them, the use of piezoelectric actuators and sensors integrated to the structure, jointly with a control system, the so-called smart structure technology, has been seen with good potentiaI. A smart structure promotes active vibration control, guaranteeing a performance increase. The objective of this work is to obtain optimal control parameters of a non-conventional vibration controller based on the fuzzy logic. A smart beam with piezoelectric actuators and sensors, that has been modeled by the finite element method, has been used to controI. The fuzzy control, which is becoming broadly utilized, mainly due to its capacity to represent complex and non-linear systems, is based in Mamdani and Takagi-Sugeno-Kang fuzzy models. The optimization scheme is based on genetic algorithms, a methodology inspired on the natural selection laws influenced by the Darwin\'s theories. Gains values and membership functions are optimized. Comparisons with the fuzzy controller achieved by trial and error parameters tuning are presented.

As características irrelevantes, presentes em bases de dados de diversos domínios, deterioram a acurácia de predição de classificadores induzidos por algoritmos de aprendizado de máquina. As bases de dados geradas por uma língua eletrônica são exemplos típicos onde a demasiada quantidade de características irrelevantes e redundantes prejudicam a acurácia dos classificadores induzidos. Para lidar com este problema, duas abordagens podem ser utilizadas. A primeira é a utilização de métodos para seleção de subconjuntos de características. A segunda abordagem é por meio de ensemble de classificadores. Um ensemble deve ser constituído por classificadores diversos e acurados. Uma forma efetiva para construção de ensembles de classificadores é por meio de seleção de características. A seleção de características para ensemble tem o objetivo adicional de encontrar subconjuntos de características que promovam acurácia e diversidade de predição nos classificadores do ensemble. Algoritmos genéticos são técnicas promissoras para seleção de características para ensemble. No entanto, a busca genética, assim como outras estratégias de busca, geralmente visam somente a construção do ensemble, permitindo que todas as características (relevantes, irrelevantes e redundantes) sejam utilizadas. Este trabalho apresenta uma abordagem baseada em algoritmos genéticos para construção de ensembles de redes neurais artificiais com um conjunto reduzido das características totais. Para melhorar a acurácia dos ensembles, duas abordagens diferenciadas para treinamento de redes neurais foram utilizadas. A primeira baseada na interrupção precoce do treinamento com o algoritmo back-propagation e a segunda baseada em otimização multi-objetivo. Os resultados obtidos comprovam a eficácia do algoritmo proposto para construção de ensembles de redes neurais acurados. Também foi constatada sua eficiência na redução das características totais, comprovando que o algoritmo proposto é capaz de construir um ensemble utilizando um conjunto reduzido de características. / The irrelevant features in databases of some domains spoil the accuracy of the classifiers induced by machine learning algorithms. Databases generated by an electronic tongue are examples where the huge quantity of irrelevant and redundant features spoils the accuracy of classifiers. There are basically two approaches to deal with this problem: feature subset selection and ensemble of classifiers. A good ensemble is composed by accurate and diverse classifiers. An effective way to construct ensembles of classifiers is to make it through feature selection. The ensemble feature selection has an additional objective: to find feature subsets to promote accuracy and diversity in the ensemble of classifiers. Genetic algorithms are promising techniques for ensemble feature selection. However, genetic search, as well as other search strategies, only aims the ensemble construction, allowing the selection of all features (relevant, irrelevant and redundant). This work proposes an approach based on genetic algorithm to construct ensembles of neural networks using a reduced feature subset of totality. Two approaches were used to train neural networks to improve the ensembles accuracy. The first is based on early stopping with back-propagation algorithm and the second is based on multi-objective optimization. The results show the effectiveness and accuracy of the proposed algorithm to construct ensembles of neural networks, and also, its efficiency in the reduction of total features was evidenced, proving its capacity for constructing an ensemble using a reduced feature subset.

O trânsito das metrópoles do novo milênio é um dos problemas que mais leva ao desperdício de recursos, com a consequente redução da qualidade de vida nas metrópoles. Os gastos ligados a vários fatores fazem com que o uso do veículo na hora do pico de trânsito seja duas vezes maior que o normal. O trânsito também gera poluição e consequentemente contribui para doenças pulmonares. O problema de planejamento operacional das infraestruturas viárias em uma grande metrópole constitui algo muito complexo. Problemas deste tamanho ainda não podem ser enfrentados, pelos sistemas computacionais modernos, na sua totalidade. Este problema se resolve dividindo as metrópoles em áreas nas quais é possível conduzir estudos que resolvam as situações locais. A circulação entre zonas distantes das metrópoles é suportada pelas vias expressas, as quais podem ser otimizadas globalmente. Com o crescimento das metrópoles e de seu número de veículos, muitas vezes, as infraestruturas ficam inadequadas, fazendo com que parte dos fluxos das vias expressas invadam os bairros. Neste trabalho se propõe fortalecer a capacidade dos bairros de enfrentar as situações de fluxos intensos. O processo proposto é hierárquico tendo uma primeira fase composta de estudos locais efetuados com simulação micro/mesoscópica, e uma sucessiva otimização global baseada nos resultados das locais. O sistema de otimização necessita de um teste que avalia as soluções escolhidas ao longo do processo. Na literatura da engenharia de tráfego se encontram diferentes níveis de análise do trânsito que geram as três categorias de modelos de simulação. Estas categorias ganham os nomes de modelos microscópicos, mesoscópicos e macroscópicos, mas nenhuma se mostra apta a ser utilizada como teste do sistema proposto. Neste trabalho se propõe um simulador que abstrai o conceito de rede viária reduzindo os custos computacionais até conseguir simular uma inteira metrópole. A técnica de estudo proposta, nos testes, se revela útil em determinadas situações, mas ainda deve ser confrontada com as novas tecnologias capazes de refinar os planos operacionais em tempo real na base dos dados de sensores e câmeras espalhados nas infraestruturas. / The traffic of new-millennium metropolises is one of the problems that most cause resources waste, consequently reducing the quality of life in these metropolises. The costs related to a series of causes make the use of vehicles at rush times be twice as frequent as during other times. The traffic also generates pollution, hence contributing to pulmonary diseases. The infrastructures operational planning problems in big cities is a complex issue. Such big problems still cannot be fully faced by modern computer systems. This can be solved dividing the cities into areas where it is possible to run studies to solve local situations. The circulation between distant areas in metropolises can be done via express motorways, which can be globally improved. With the expansion of big cities and their vehicles, the infrastructures frequently become inadequate and the stream invades neighborhoods. The aim of the present work is to improve the capacity of neighborhoods streams, offering besides technical norms, a global optimization based on local results. The optimization system needs a test that evaluates the chosen solutions along the process. In traffic engineering literature, there are different levels of traffic analysis that generate the three simulation model categories. These categories are named microscopic, mesoscopic, and macroscopic models, but none of them is able to be used as test to the proposed system. In this work, it is proposed a simulator that abstracts the concept of road network, reducing the computer expenses up to the simulation of a whole city. The study technic pruioposed in the tests shows itself as useful in certain situations, but still must be confronted with new technologies able to refine the operational plans in real time based on the sensors and cameras data.

Submitted by Cássia Santos (cassia.bcufg@gmail.com) on 2014-09-19T11:19:07Z No. of bitstreams: 2 Dissertacao Daniel Vitor de Lucena.pdf: 708978 bytes, checksum: 466a21a76649073c30364b80f17037fc (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2014-09-19T11:25:02Z (GMT) No. of bitstreams: 2 Dissertacao Daniel Vitor de Lucena.pdf: 708978 bytes, checksum: 466a21a76649073c30364b80f17037fc (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2014-09-19T11:25:02Z (GMT). No. of bitstreams: 2 Dissertacao Daniel Vitor de Lucena.pdf: 708978 bytes, checksum: 466a21a76649073c30364b80f17037fc (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2013-05-03 / This work proposes the use of multi-objective genetics algorithms NSGA-II and SPEA-II on the variable selection in multivariate calibration problems. These algorithms are used for selecting variables for a Multiple Linear Regression (MLR) by two conflicting objectives: the prediction error and the used variables number in MLR. For the case study are used wheat data obtained by NIR spectrometry with the objective for determining a variable subgroup with information about protein concentration. The results of traditional techniques of multivariate calibration as the Partial Least Square (PLS) and Successive Projection Algorithm (SPA) for MLR are presents for comparisons. The obtained results showed that the proposed approach obtained better results when compared with a monoobjective evolutionary algorithm and with traditional techniques of multivariate calibration. / Este trabalho propõe a utilização dos algoritmos genéticos multiobjetivo NSGA-II e SPEA-II na seleção de variáveis em problemas de calibração multivariada. Esses algoritmos são utilizados para selecionar variáveis para Regressão Linear Múltipla (MLR) com dois objetivos conflitantes: o erro de predição e do número de variáveis utilizadas na MLR. Para o estudo de caso são usado dados de trigo obtidos por espectrometria NIR com o objetivo de determinar um subgrupo de variáveis com informações sobre a concentração de proteína. Os resultados das técnicas tradicionais de calibração multivariada como dos Mínimos Quadrados Parciais (PLS) e Algoritmo de Projeções Sucessivas (APS) para a MLR estão presentes para comparações. Os resultados obtidos mostraram que a abordagem proposta obteve melhores resultados quando comparado com um algoritmo evolutivo monoobjetivo e com as técnicas tradicionais de calibração multivariada.

L’objectif de ce travail est de développer un système d’affectation des conteneurs aux véhicules autonomes intelligents (AIVs) dans un terminal à conteneurs. Dans la première phase, on a développé un système statique pour résoudre le problème multi-objectif optimisant la durée totale des opérations de déplacement des conteneurs, le temps d’attente des véhicules aux niveaux de points de chargement et de déchargement et l’équilibre de temps de travail entre les véhicules. L’approche proposée est l’algorithme génétique(AG). Une extension de cette approche a été ensuite effectuée pour corriger les limites de la précédente. Pour choisir la meilleure approche, une étude comparative a été réalisée entre trois approches : AG, AG & DIJK et AG & DIJK & HEUR. Les résultats numérique ont montré que l’approche AG & DIJK & HEUR est meilleure. Dans la deuxième phase, on a étudié la robustesse de notre système dans un environnement dynamique. Un retard de l’arrivée d’un navire au port ou un dysfonctionnement de l’un des équipements peutperturber le planning des opérations et donc influencer sur les opérations d’affectation des conteneurs. L’idée était d’ajouter les nouveaux conteneurs aux véhicules qui sont déjà non disponibles. D’autres cas de perturbation comme la congestion routière, la non disponibilité de certaines portions de la routes ont été étudiés expérimentalementEt les résultats numériques ont montré la robustesse de notre approche pour le cas dynamique.Mots-clés : Conteneurs, AIV, routage, optimisation, algorithme génetique, environnement dynamique. / The objective of our work is to develop a container assignment system for intelligent autonomous vehicles (AIVS) in a container terminal. Given the complexity of this problem, it was proposed to decompose it into three problems: The problem of dispatching containers to AIVS, the AIVS routing problem and the problem of scheduling containers to queues of AIVS. To achieve this goal, we developed in the first phase, a static system for multi-objective problem to optimize the total duration of the containers transportation, the waiting time of vehicles at loading points and the equilibrium of working time between vehicles. The approach used was the genetic algorithm (GA). This approach was applied to optimize only the assignment operation without influence on the choice of the path traveled by each AIV. An extension of this work was then made to improve the results found. For this purpose, a comparative study was carried out between three approaches: The first approach is the AG, the second approach is the GA and the Dijkstra algorithm (DIJK) that was used to find the shortest path for each vehicle and the third approach is the AG and DIJK and heuristic (HEUR) which was proposed to choose the nearest vehicle of each container. The numerical study showed the best performance of the AG & DJK & HEUR approach over the other two approaches. In the second phase of our project, the robustness of our system in a dynamic environment has been studied. A delay of the arrival of a ship at the port or malfunction of one of any equipment of the port can cause a delay of one of the operations of loading or unloading process. This will affect the container assignment operation. The idea was to add new containers to vehicles that are already unavailable. The traffic can also cause a delay in arrival of the vehicle at the position of the container or the unavailability of one of the paths crossing point. These cases were investigated experimentally, numerical results showed the robustness of our approach to dynamic case.

[pt] Em muitas aplicações reais de processamento de sinais, as observações do fenômeno em estudo chegam seqüencialmente no tempo. Consequentemente, a tarefa de análise destes dados envolve estimar quantidades desconhecidas em cada observação concebida do fenômeno. Na maioria destas aplicações, entretanto, algum conhecimento prévio sobre o fenômeno a ser modelado está disponível. Este conhecimento prévio permite formular modelos Bayesianos, isto é, uma distribuição a priori sobre as quantidades desconhecidas e uma função de verossimilhança relacionando estas quantidades com as observações do fenômeno. Dentro desta configuração, a inferência Bayesiana das quantidades desconhecidas é baseada na distribuição a posteriori, que é obtida através do teorema de Bayes. Infelizmente, nem sempre é possível obter uma solução analítica exata para esta distribuição a posteriori. Graças ao advento de um formidável poder computacional a baixo custo, em conjunto com os recentes desenvolvimentos na área de simulações estocásticas, este problema tem sido superado, uma vez que esta distribuição a posteriori pode ser aproximada numericamente através de uma distribuição discreta, formada por um conjunto de amostras. Neste contexto, este trabalho aborda o campo de simulações estocásticas sob a ótica da genética Mendeliana e do princípio evolucionário da sobrevivência dos mais aptos. Neste enfoque, o conjunto de amostras que aproxima a distribuição a posteriori pode ser visto como uma população de indivíduos que tentam sobreviver num ambiente Darwiniano, sendo o indivíduo mais forte, aquele que possui maior probabilidade. Com base nesta analogia, introduziu-se na área de simulações estocásticas (a) novas definições de núcleos de transição inspirados nos operadores genéticos de cruzamento e mutação e (b) novas definições para a probabilidade de aceitação, inspirados no esquema de seleção, presente nos Algoritmos Genéticos. Como contribuição deste trabalho está o estabelecimento de uma equivalência entre o teorema de Bayes e o princípio evolucionário, permitindo, assim, o desenvolvimento de um novo mecanismo de busca da solução ótima das quantidades desconhecidas, denominado de inferência evolucionária. Destacamse também: (a) o desenvolvimento do Filtro de Partículas Genéticas, que é um algoritmo de aprendizado online e (b) o Filtro Evolutivo, que é um algoritmo de aprendizado batch. Além disso, mostra-se que o Filtro Evolutivo, é em essência um Algoritmo Genético pois, além da sua capacidade de convergência a distribuições de probabilidade, o Filtro Evolutivo converge também a sua moda global. Em conseqüência, a fundamentação teórica do Filtro Evolutivo demonstra, analiticamente, a convergência dos Algoritmos Genéticos em espaços contínuos. Com base na análise teórica de convergência dos algoritmos de aprendizado baseados na inferência evolucionária e nos resultados dos experimentos numéricos, comprova-se que esta abordagem se aplica a problemas reais de processamento de sinais, uma vez que permite analisar sinais complexos caracterizados por comportamentos não-lineares, não- gaussianos e nãoestacionários. / [en] In many real-world signal processing applications, the phenomenon s observations arrive sequentially in time; consequently, the signal data analysis task involves estimating unknown quantities for each phenomenon observation. However, in most of these applications, prior knowledge about the phenomenon being modeled is available. This prior knowledge allows us to formulate a Bayesian model, which is a prior distribution for the unknown quantities and the likelihood functions relating these quantities to the observations. Within these settings, the Bayesian inference on the unknown quantities is based on the posterior distributions obtained from the Bayes theorem. Unfortunately, it is not always possible to obtain a closed-form analytical solution for this posterior distribution. By the advent of a cheap and formidable computational power, in conjunction with some recent developments in stochastic simulations, this problem has been overcome, since this posterior distribution can be obtained by numerical approximation. Within this context, this work studies the stochastic simulation field from the Mendelian genetic view, as well as the evolutionary principle of the survival of the fittest perspective. In this approach, the set of samples that approximate the posteriori distribution can be seen as a population of individuals which are trying to survival in a Darwinian environment, where the strongest individual is the one with the highest probability. Based in this analogy, we introduce into the stochastic simulation field: (a) new definitions for the transition kernel, inspired in the genetic operators of crossover and mutation and (b) new definitions for the acceptation probability, inspired in the selection scheme used in the Genetic Algorithms. The contribution of this work is the establishment of a relation between the Bayes theorem and the evolutionary principle, allowing the development of a new optimal solution search engine for the unknown quantities, called evolutionary inference. Other contributions: (a) the development of the Genetic Particle Filter, which is an evolutionary online learning algorithm and (b) the Evolution Filter, which is an evolutionary batch learning algorithm. Moreover, we show that the Evolution Filter is a Genetic algorithm, since, besides its capacity of convergence to probability distributions, it also converges to its global modal distribution. As a consequence, the theoretical foundation of the Evolution Filter demonstrates the convergence of Genetic Algorithms in continuous search space. Through the theoretical convergence analysis of the learning algorithms based on the evolutionary inference, as well as the numerical experiments results, we verify that this approach can be applied to real problems of signal processing, since it allows us to analyze complex signals characterized by non-linear, nongaussian and non-stationary behaviors.

The assumption in economics of a representative agent is often made. However, it is a very rigid assumption. Hall and Jones (2004b) presented an economic model that essentially provided for a representative agent for each age group in determining the group's health level function. Our work seeks to extend their theoretical version of the model by allowing for two representative agents for each age—one for each of “Healthy” and “Sick” risk-factor groups—to allow for additional heterogeneity in the populace. The approach to include even more risk-factor groups is also briefly discussed. While our “extended” theoretical model is not applied directly to relevant data, several techniques that could be applicable were the relevant data to be obtained are demonstrated on other data sets. This includes examples of using linear classification, fitting baseline-category logit models, and running the genetic algorithm.

Procedure for site discovery, system design, and optimization of power harvesting systems is developed with an emphasis on application to air compressors. Limitations for the usage of infrared pyrometers is evaluated. A system of governing equations for thermoelectric generators is developed. A solution method for solving the system of equations is created in order to predict power output from the device. Payback analysis is proposed for determining economic viability. A genetic algorithm is used to optimize the power harvesting system payback with changing quantities and varieties of thermoelectric generators, as well as the back work put into cooling the thermoelectric generators. Experimental data is taken for laboratory simulation of a power harvesting system under varying resistive load and thermal conductances in order to confirm the working model. A power harvester is designed for and installed on a consumer grade portable air compressor. Experimental data is compared against the model's prediction. As a case study, a system is designed for a water-cooled power harvesting system. Thermoelectric generator power harvesters are found to be economically infeasible for typical installations at current energy prices. Changes in parameters which would increase economic feasibility of the power harvesting system are discussed.

Procedure for site discovery, system design, and optimization of power harvesting systems is developed with an emphasis on application to air compressors. Limitations for the usage of infrared pyrometers is evaluated. A system of governing equations for thermoelectric generators is developed. A solution method for solving the system of equations is created in order to predict power output from the device. Payback analysis is proposed for determining economic viability. A genetic algorithm is used to optimize the power harvesting system payback with changing quantities and varieties of thermoelectric generators, as well as the back work put into cooling the thermoelectric generators.Experimental data is taken for laboratory simulation of a power harvesting system under varying resistive load and thermal conductances in order to confirm the working model. A power harvester is designed for and installed on a consumer grade portable air compressor. Experimental data is compared against the model's prediction. As a case study, a system is designed for a water-cooled power harvesting system.Thermoelectric generator power harvesters are found to be economically infeasible for typical installations at current energy prices. Changes in parameters which would increase economic feasibility of the power harvesting system are discussed.

<div>Multi-model inference covers a wide range of modern statistical applications such as variable selection, model confidence set, model averaging and variable importance.</div><div>The performance of multi-model inference depends on the availability of candidate models, whose quality has been rarely studied in literature. In this dissertation, we study genetic algorithm (GA) in order to obtain high-quality candidate models. Inspired by the process of natural selection, GA performs genetic operations such as selection, crossover and mutation iteratively to update a collection of potential solutions (models) until convergence. The convergence properties are studied based on the Markov chain theory and used to design an adaptive termination criterion that vastly reduces the computational cost. In addition, a new schema theory is established to characterize how the current model set is improved through evolutionary process. Extensive numerical experiments are carried out to verify our theory and demonstrate the empirical power of GA, and new findings are obtained for two real data examples. </div>