Thesis (MScEng)--Stellenbosch University, 2011. / ENGLISH ABSTRACT: Reconfigurable Manufacturing Systems (RMSs) have the ability to reconfigure hardware and control resources at all of the functional and organizational levels. This allows for quick adjustment of production capacity and functionality in response to sudden changes in market or in regulatory requirements. This study evaluates the characteristics and operation of automated reconfigurable assembly lines using discrete event simulation. The assembly line uses a conveyor system which transports pallets to various machines to perform the assembly process. Different conveyor configurations are developed for the same assembly process using Simio simulation software. A part family consisting of five variants are assembled on the same assembly line with a large variation in the production quantities for each product. This requires the assembly system to be able to quickly adjust its functionality and capacity. Multi-objective optimization is performed on the models through the use of a Pareto exhaustive search experiment. The two contradicting objectives used are the throughput rate of the system and the average work in progress, with the aim of maximizing the former and minimizing the latter. From the Pareto exhaustive search experiment, a Pareto front is constructed showing which configuration is preferred under certain operation conditions. However it is concluded that the Pareto front can be tailored to fit the specific needs of the decision maker, depending on what the decision maker is willing to pay. An experiment that evaluates the effect of changing the conveyor speed is performed. It is established that under certain operating conditions, increasing the conveyor speed higher than the ceiling value will not improve the performance of the system. A production scenario was also developed which include different order sizes for each of the five parts of the part family. The configurations have to alter their capacities based on the order sizes to test which system performs the best under these operating conditions. For this experiment, the ramp-up time was of interest but the best system was chosen based on the combination of throughput rate and the average work in progress. From the results of the different experiments, it is recommended to first determine the maximum capacity and the operating logic before choosing one of the configurations. Once this is decided, the information gathered from the experiments can then be tailored for the decision maker to establish the best operating conditions for the chosen con guration. The developed simulation models are used as a Decision Support System for future research on the topic. It is recommended for future research to focus on using Automated Guided Vehicles (AGVs) instead of a conveyor system as transportation method. / AFRIKAANSE OPSOMMING: Herkonfigureerbare Vervaardigingstelsels (HVSs) het die vermoee om alle hardeware en beheer hulpbronne, op alle funksionele en organisatoriese vlakke te herkonfigureer. Dit maak dit moontlik vir vinnige verstellings aan produksie kapasiteit en funksionaliteit, indien daar 'n skielike verandering in die mark of wetgewing is. Hierdie studie evalueer die karakteristieke en werking van outomatiese herkon gureerbare monteerlyne met behulp van diskrete gebeurtenis simulasie. Die monteerlyne gebruik vervoerbande om pallette na verskeie masjiene te vervoer, sodat die parte aanmekaar gesit kan word. Simio simulasie sagteware is gebruik om verskillende vervoerband kon gurasies vir dieselfde monteringsproses te ontwikkel. 'n Part familie van vyf variante word op dieselfde monteerlyn aanmekaargesit. Daar is 'n groot variasie in die produksie hoeveelhede van elk van die vyf variante, dus moet die monteerlyne vinnig die kapasiteit en funksionaliteit kan aanpas. Multi-doelwitoptimering is toegepas op die modelle deur 'n Pareto alomvattende soek eksperiment uit te voer. Die twee teenstrydige doelwitte wat gebruik is, is die deurset tempo van die stelsel asook die gemiddelde werk-in-proses. Die doel is om die deurset tempo te maksimeer en terselfde tyd die gemiddelde werk-in-proses te minimeer. Die Pareto alomvattende soek eksperiment word verder gebruik om 'n Pareto front te skep wat uitwys watter vervoerband kon gurasies verkies word onder sekere bedryfstoestande. Die Pareto front kan egter aangepas word om die spesi eke behoeftes van die besluitnemer te pas. 'n Eksperiment is uitgevoer om die uitwerking van die vervoerbandspoed op die stelsel te toets. Resultate het getoon dat onder sekere bedryfstoestande die werkverrigting van die stelsel nie verbeter indien die spoed 'n maksimum grenswaarde oorskry nie. 'n Eksperiment wat 'n produksie scenario voorstel is ontwikkel waarin die vraag na die vyf part variante gevarieer word. Die vervoerband konfigurasies moet dan die kapasiteit aanpas gebaseer op die vraag na die parte. Die doel van die eksperiment is om te toets watter kon gurasie die beste vaar onder hierdie bedryfstoestande. Die tyd wat dit neem vir die stelsel om weer op dreef te kom na 'n verandering in kapasiteit is ondersoek in hierdie eksperiment, maar die beste stelsel is nog steeds gekies gebaseer op die kombinasie van deurset tempo en die gemiddelde werk-in-proses. Gegewe die resultate van die verskillende eksperimente, word dit voorgestel dat die besluitnemer eers die maksimum kapasiteit en die bedryfstoestande vasstel, voordat 'n vervoerband kon gurasie gekies word. Sodra dit besluit is, kan die inligting wat tydens die eksperimente ingesamel is, aangepas word om die beste bedryfstoestande vir die kon gurasie wat gekies is, vas te stel. Die simulasie modelle wat ontwikkel is word gebruik as 'n besluitnemingsondersteuningstelsel vir toekomstige navorsing oor die onderwerp. Dit word voorgestel dat toekomstige navorsing die moontlikheid van geoutomatiseerde begeleide voertuie (GBV), in plaas van vervoerbande as vervoermiddel, ondersoek.

Dans ce travail, un moteur Stirling de type Gamma alimenté par énergie solaire avec une faible différence de température a été étudié numériquement et expérimentalement. Un nouveau modèle appelé Polytropic Stirling Model with Losses (PSML) a été proposé et appliqué au moteur GPU-3 Stirling. Un cryoréfrigérateur basé sur un moteur Stirling intégral de type Alpha a été étudié numériquement, après avoir mesuré ses dimensions géométriques au laboratoire. Pour le moteur Stirling de type gamma du laboratoire, le modèle ait thermodynamique à vitesse finie et le modèle isotherme a été développé, incluant les bilans de masse et d’énergie à travers les différents volumes (compression, régénération et expansion) dans le moteur. Différents types de pertes thermiques et mécaniques ont été considérés dans le modèle afin d'analyser les processus thermodynamiques et les pertes dans le moteur Stirling. En outre, des études paramétriques sur les performances du moteur Stirling alimenté à l’énergie solaire ont également été étudiées expérimentalement et numériquement. La comparaison entre les résultats expérimentaux et les résultats de simulation à différents déphasages entre le déplaceur et le piston, et à différentes course de piston montre que le modèle est convaincant dans la prédiction des performances du moteur Stirling. Basé sur la méthode thermodynamique en dimension physique finie, une méthode d’algorithme génétique multi-objectives, objectifs étant la puissance fournie, le rendement énergétique et le taux de génération d'entropie a été utilisé pour optimiser la fonction et la géométrie du moteur du type Gamma. En comparant avec la méthode d'optimisation écologique, la méthode multi-objectif permet de mieux équilibrer les trois objectifs. Le nouveau modèle (PSML) proposé pour prédire les performances du moteur de type Bêta ou Gamma du moteur Stirling, il divise l'espace de travail en 5 parties (volume de compression, refroidisseur, régénérateur, chauffage et volume d'extension). Une liaison entre volume de compression et volume d'extension a été ajoutée dans le modèle adiabatique classique du moteur Stirling. Ainsi, des processus polytropiques ont été considérés dans les volumes de compression et d'expansion du moteur Stirling. Le moteur Stirling GPU-3 a été utilisé pour valider le nouveau modèle. Il a été démontré que le nouveau modèle (PSML) prédit correctement la puissance de sortie et le rendement du moteur. Dans la dernière partie de la thèse, un Cryorefroidisseur Stirling de type Alpha, a été étudié en utilisant un modèle isotherme prenant en considération différentes pertes. Les volumes de compression et d'expansion sont considérés isothermes, et la variation de la température du régénérateur est considéré linéaire. Les bilans d'énergie et d'exergie du Cryorefroidisseur ont été réalisés, et l'effet de divers paramètres sur la performance (puissance de refroidissement et puissance mécanique consommée) est étudié. Les résultats de la simulation pour PPG-102 Stirling cryocooler ont été comparés avec deux autres résultats de simulation de la littérature et des résultats expérimentaux indiquant que ce modèle est convaincant pour prédire la performance du Cryorefroidisseur. / In this work a solar powered low temperature difference Gamma type Stirling engine has been studied experimentally and numerically using an isothermal model coupled with various losses and using an objective optimization. A new model named Polytropic Stirling Model with Losses (PSML) has been proposed which was applied to the Beta type GPU-3 Stirling engine. An Alpha type integral Stirling cryocooler has been studied numerically using an isothermal model with losses. To study a Gamma type Stirling engine of our laboratory, an isothermal model coupled with finite speed method has been developed, including mass and energy balances through different spaces of the engine. The engine is divided into 3 volumes: compression volume, regeneration volume, and expansion volume. Different kind of thermal and mechanical losses have been considered in the model, in order to analyze thermodynamic processes and losses in the Stirling Engine. In addition, parameter effects on the performance of the solar powered gamma type Stirling engine have also been studied experimentally and numerically. The comparison between the experimental results and the simulation results at different phase shift between the displacer and the piston, and at different piston stroke shows that the model is convincing to predict the Stirling engine performance. Based on the Finite Physical Dimensional Thermodynamic method, a multi-objective genetic method considering output power, thermal efficiency and entropy generating rate as objective functions simultaneously, has been used to multi-objective optimize the Gamma type Stirling engine. Comparing with the ecological optimization method, the multi-objective method can better balance the three objective goals. The new model (PSML) proposed in the thesis for predicting performance of Beta or Gamma type of Stirling engine divides the working space into 5 parts (compression volume, cooler, regenerator, heater, and expansion volume). A bypass linking compression volume and expansion volume has been added in the classic adiabatic model of Stirling engine. Thus, polytropic processes have been considered in the compression and expansion volumes of the Stirling engine. The GPU-3 Stirling engine has been used to validate the new model. It was shown that the new model (PSML) predict well the output power and the thermal efficiency of the engine well. An isothermal model considering various losses was developed and presented in the last part of this thesis to study an Alpha type Stirling cryocooler, whose geometrical dimensions were measured in our laboratory. The compression and expansion volumes are supposed to be isothermal, the variation of the regenerator temperature is supposed to be linear. Energy and exergy balances of the cryocooler were developed. The effect of various parameters on the cryocooler performance (cooling power and input power) are investigated. The simulation results for PPG-102 Stirling cryocooler were compared with two other simulation results of the literature and with experimental results which indicated that this model is convincing to predict the performance of the Stirling cyocooler

Because of the increasing demand for the provision of energy worldwide and the numerous damages caused by a major use of fossil sources, the contribution of renewable energies has been increasing significantly in the global energy mix with the aim at moving towards a more sustainable development. In this context, this work aims at the development of a general methodology for designing PV systems based on ecodesign principles and taking into account simultaneously both techno-economic and environmental considerations. In order to evaluate the environmental performance of PV systems, an environmental assessment technique was used based on Life Cycle Assessment (LCA). The environmental model was successfully coupled with the design stage model of a PV grid-connected system (PVGCS). The PVGCS design model was then developed involving the estimation of solar radiation received in a specific geographic location, the calculation of the annual energy generated from the solar radiation received, the characteristics of the different components and the evaluation of the techno-economic criteria through Energy PayBack Time (EPBT) and PayBack Time (PBT). The performance model was then embedded in an outer multi-objective genetic algorithm optimization loop based on a variant of NSGA-II. A set of Pareto solutions was generated representing the optimal trade-off between the objectives considered in the analysis. A multi-variable statistical method (i.e., Principal Componet Analysis, PCA) was then applied to detect and omit redundant objectives that could be left out of the analysis without disturbing the main features of the solution space. Finally, a decision-making tool based on M-TOPSIS was used to select the alternative that provided a better compromise among all the objective functions that have been investigated. The results showed that while the PV modules based on c-Si have a better performance in energy generation, the environmental aspect is what makes them fall to the last positions. TF PV modules present the best trade-off in all scenarios under consideration. A special attention was paid to recycling process of PV module even if there is not yet enough information currently available for all the technologies evaluated. The main cause of this lack of information is the lifetime of PV modules. The data relative to the recycling processes for m-Si and CdTe PV technologies were introduced in the optimization procedure for ecodesign. By considering energy production and EPBT as optimization criteria into a bi-objective optimization cases, the importance of the benefits of PV modules end-of-life management was confirmed. An economic study of the recycling strategy must be investigated in order to have a more comprehensive view for decision making.

O presente trabalho otimizou os parâmetros concentrados do sistema de suspensão de um modelo de veículo completo, representando um automóvel de passeio que trafega a uma velocidade constante por um determinado perfil de pista previsto na norma ISO 8608, 1995, através da utilização de um algoritmo meta-heurístico de otimização multiobjetivo. Duas rotinas numérico-computacionais foram desenvolvidas, visando realizar tal otimização tanto no domínio do tempo quanto no domínio da frequência. A utilização de algoritmos meta-heurísticos vem ganhando espaço na otimização de sistemas mecânicos, proporcionando rapidez e precisão na obtenção de resultados ótimos. Ao se combinar um algoritmo de otimização a um modelo que represente satisfatoriamente um sistema mecânico, obtém-se uma ferramenta indicadora dos parâmetros de máxima eficiência do sistema, que pode ser utilizada em inúmeras aplicações. Pretendeu-se, com a integração de rotinas de análise dinâmica nos domínios do tempo e da frequência ao algoritmo genético de otimização multiobjetivo NSGA-II, desenvolvido por Deb et al., 2002, a obtenção de duas fronteiras ótimas de Pareto. Estas fronteiras consistem no conjunto de soluções não dominadas que minimizam as seguintes funções objetivo: o valor RMS ponderado da aceleração vertical do assento do motorista, o valor RMS da média do fator de amplificação dinâmica das quatro rodas do modelo e o máximo deslocamento relativo entre cada roda e a carroceria. O método proposto por Shinozuka e Jan, 1972, é utilizado para a obtenção do perfil de irregularidades da pista no domínio do tempo a partir das equações de densidade espectral de potência (PSD) que representam as diferentes classes de pavimentos. O método de Newmark, 1959, é utilizado para resolver a equação diferencial de movimento no domínio do tempo e obter a resposta dinâmica do modelo a tais irregularidades. O comportamento dinâmico do modelo de veículo no domínio da frequência foi obtido através da utilização da função de resposta em frequência (FRF) do modelo de veículo analisado. Os resultados demonstraram a capacidade de ambas as rotinas de análise dinâmica desenvolvidas de produzir resultados consistentes com os encontrados na literatura, bem como a capacidade dos algoritmos de otimização implementados de fornecer fronteiras ótimas de Pareto para os problemas propostos. / The proposed work optimized the concentrated parameters of a full-vehicle model’s suspension system, being that model representative of a passenger car which travels at a constant speed on a certain road profile provided by the ISO 8608, 1995, standard, using a multi-objective meta-heuristic optimization algorithm. Two numerical-computational routines were developed, seeking to perform said optimization for both the time and frequency domains. The use of meta-heuristic algorithms has been increasing in mechanical systems optimization, providing speed and accuracy in obtaining an optimal result. Combining an optimization algorithm with a model that satisfactorily represents a mechanical system yields a tool that indicates the system’s maximum efficiency parameters, which can be used in numerous applications. It was intended, with the integration of the dynamic analysis routines to the multi-objective genetic optimization algorithm NSGA-II, developed by Deb et al., 2002, the obtainment of two Pareto-optimal fronts. These fronts consist in the set of non-dominated solutions that minimize the following objective functions: the weighted RMS value of the driver’s seat vertical acceleration, the mean RMS value of the model wheel’s dynamic amplification factor, and the maximum relative displacement between each wheel and the body of the vehicle model. The method proposed by Shinozuka and Jan, 1972, is used to obtain the road irregularity profile in the time domain from the power spectral density (PSD) equations that represent the different pavement classes. The Newmark’s method (1959) is used to solve the differential motion equation in the time domain, in order to obtain the vehicle model’s responses to these irregularities. The dynamic behavior of the vehicle model in the frequency domain was obtained through the use of the frequency response function (FRF) of the analyzed model. The results showed the capacity of both the dynamic analysis routines developed in generating results that are consistent with those found in literature, as well as the capacity of the optimization algorithms implemented in providing Pareto optimal fronts to the proposed problems.

L'informatique a changé profondément les aspects méthodologiques du processus de découverte dans les différents domaines du savoir. Les chercheurs ont à leur disposition aujourd'hui de nouvelles capacités qui permettent d'envisager la résolution de nouveaux problèmes. Les plates-formes parallèles et distribués composées de ressources partagés entre différents participants peuvent rendre ces nouvelles capacités accessibles à tout chercheur et offre une puissance de calcul qui a été limitée jusqu'à présent, aux projets scientifiques les plus grands (et les plus riches). Dans ce document qui regroupe les résultats obtenus pendant mon doctorat, nous explorons quatre facettes différentes de la façon dont les organisations s'engagent dans une collaboration sur de plates-formes parallèles et distribuées. En utilisant des outils classiques de l'analyse combinatoire, de l'ordonnancement multi-objectif et de la théorie des jeux, nous avons montré comment calculer des ordonnancements avec un bon compromis entre les résultats obtenu par les participants et la performance globale de la plate-forme. En assurant des résultats justes et en garantissant des améliorations de performance pour les différents participants, nous pouvons créer une plate-forme efficace où chacun se sent toujours encourager à collaborer et à partager ses ressources. Tout d'abord, nous étudions la collaboration entre organisations égoïstes. Nous montrons que le comportement égoïste entre les participants impose une borne inférieure sur le makespan global. Nous présentons des algorithmes qui font face à l'égoïsme des organisations et qui présentent des résultats équitables. La seconde étude porte sur la collaboration entre les organisations qui peuvent tolérer une dégradation limitée de leur performance si cela peut aider à améliorer le makespan global. Nous améliorons les bornes d'inapproximabilité connues sur ce problème et nous présentons de nouveaux algorithmes dont les garanties sont proches de l'ensemble de Pareto (qui regroupe les meilleures solutions possibles). La troisième forme de collaboration étudiée est celle entre des participants rationnels qui peuvent choisir la meilleure stratégie pour leur tâches. Nous présentons un modèle de jeu non coopératif pour le problème et nous montrons comment l'utilisation de "coordination mechanisms" permet la création d'équilibres approchés avec un prix de l'anarchie borné. Finalement, nous étudions la collaboration entre utilisateurs partageant un ensemble de ressources communes. Nous présentons une méthode qui énumère la frontière des solutions avec des meilleurs compromis pour les utilisateurs et sélectionne la solution qui apporte la meilleure performance globale. / Computer science is deeply changing methodological aspects of the discovery process in different areas of knowledge. Researchers have at their disposal new capabilities that can create novel research opportunities. Parallel and distributed platforms composed of resources shared between different participants can make these new capabilities accessible to every researcher at every level, delivering computational power that was restricted before to bigger (and wealthy) scientific projects. This work explores four different facets of the rules that govern how organizations engage in collaboration on modern parallel and distributed platforms. Using classical combinatorial tools, multi-objective scheduling and game-theory, we showed how to compute schedules with good trade-offs between the results got by the participants and the global performance of the platform. By ensuring fair results and guaranteeing performance improvements for the participants, we can create an efficient platform where everyone always feels encouraged to collaborate and to share its resources. First, we study the collaboration between selfish organizations. We show how the selfish behavior between the participants imposes a lower bound on the global makespan. We present algorithms that cope with the selfishness of the organizations and that achieve good fairness in practice. The second study is about collaboration between organizations that can tolerate a limited degradation on their performance if this can help ameliorate the global makespan. We improve the existing inapproximation bounds for this problem and present new algorithms whose guarantees are close to the Pareto set. The third form of collaboration studied is between rational participants that can independently choose the best strategy for their jobs. We present a non-cooperative game-theoretic model for the problem and show how coordination mechanisms allow the creation of approximate pure equilibria with bounded price of anarchy. Finally, we study collaboration between users sharing a set of common resources. We present a method that enumerates the frontier of best compromise solutions for the users and selects the solution that brings the best value for the global performance function.

Esta dissertação apresenta os algoritmos BicriterionAnt, MACS e MONACO, disponíveis na literatura, baseados em colônia de formigas, para resolução do Problema do Caixeiro Viajante Multiobjetivo (PCVMO). São apresentadas as características do problema e de cada algoritmo utilizado. Estes algoritmos foram testados em seis instâncias bi-objetivo do PCVMO. Foram implementadas algumas alterações na estrutura original dos algoritmos na tentativa de produzir resultados melhores do que os algoritmos originais. Para a avaliação dos resultados e medição da qualidade das soluções, foram utilizadas métricas de desempenho que auxiliam na identificação dos melhores conjuntos de soluções não-dominadas. / This dissertation presents the BicriterionAnt, MACS and MONACO Ant Colony algorithms, available in literature, to solve the Multi-Objective Traveling Salesman Problem (MOTSP). The characteristics of the problem and of each algorithm used are presented. Those algorithms were tested in six bi-objective instances of MOTSP. Changes in the original algorithms were implemented to try to produce better results than the original ones. To validate the results and to measure the quality of the solutions, metrics of performance were used which help to identify the best non-dominated solution sets.

Fundação de Amparo a Pesquisa do Estado de Minas Gerais / Several studies have investigated the properties of the workspace of opened robotic chains (or serial) with the purpose of emphasizing its geometric and kinematic characteristics, to devise analytical algorithms and procedures for its design. The workspace of a robot manipulator is considered of great interest from theoretical and practical viewpoint. In classical applications in industry, manipulators need to pass through singularities in the joint space to change their posture. A 3-DOF manipulator can execute a non-singular change of posture if and only if there is at least one point in its workspace which has exactly three coincident solutions of the Inverse Kinematic Model (IKM). It is very difficult to express this condition directly from the kinematic model. Thus, in this work, the algebraic tool Gröbner basis is used to obtain an equation for splitting the regions with different types of 3R orthogonal manipulators. The determinant of Jacobian matrix of the direct kinematic model is considered equal to zero to obtain the other surfaces of separation. In addition, is presented a classification of 3R orthogonal manipulators related to the number of solutions in IKM, the number of cusp points and nodes. Some problems of multi-objective optimization are proposed to obtain the optimal design of robots. First considering a general case where the aim is to maximize the volume of the workspace, maximize the stiffness of the joint system and optimize the dexterity of the manipulator without the imposition of restrictions. Next, the optimization problem is subject to penalties that control the topology, making it possible to obtain solutions which satisfy the predetermined topologies. Solutions are presented for the case r3 null and r3 not null. The optimization problem is investigated by using a deterministic technique and two evolutionary algorithms. Some numerical applications are presented to show the efficiency of the proposed methodology. / Diversos estudos têm investigado as propriedades do espaço de trabalho de cadeias robóticas abertas com o objetivo de enfatizar suas características geométricas e cinemáticas, criar algoritmos analíticos e procedimentos para o seu projeto. O espaço de trabalho de um robô manipulador é considerado de grande interesse do ponto de vista teórico e prático. Em aplicações clássicas na indústria, manipuladores precisam passar por singularidades no espaço das juntas para mudar sua postura. Um manipulador com três graus de liberdade pode executar uma mudança de postura não singular se, e somente se, existe pelo menos um ponto em seu espaço de trabalho que tem exatamente três soluções coincidentes do Modelo Geométrico Inverso (MGI). É muito difícil expressar esta condição a partir do modelo cinemático. Assim, neste trabalho, a ferramenta algébrica base de Groebner é utilizada para obter uma das equações que separam as regiões que possuem diferentes tipos de manipuladores 3R ortogonais. O determinante da matriz Jacobiana do Modelo Geométrico Direto é considerado nulo para obter as demais superfícies de separação. Além disso, apresenta-se uma classificação dos manipuladores 3R ortogonais em relação ao número de soluções no MGI, o número de pontos de cúspides e o número de nós. Alguns problemas de otimização multi-objetivo são propostos visando obter o projeto ótimo de robôs. Primeiramente, considera-se o caso geral, cujo objetivo é maximizar o volume do espaço de trabalho, maximizar a rigidez do sistema de juntas e otimizar a destreza do manipulador sem a imposição de restrições. Em seguida, o problema de otimização é sujeito a penalidades que controlam a topologia, tornando possível a obtenção de soluções que obedeçam as topologias pré-estabelecidas. São apresentadas as soluções para o caso r3 nulo e para r3 não nulo. O problema de otimização é investigado aplicando uma técnica determinística e dois algoritmos evolutivos. Algumas aplicações numéricas são apresentadas para mostrar a eficiência da metodologia proposta. / Doutor em Engenharia Mecânica

Fundação de Amparo a Pesquisa do Estado de Minas Gerais / In the last few decades, the application of optimization in engineering problems has grown considerably. There are various optimization methods and the performance of each one depends on the type of problem considered. The natural methods, which are based on probabilistic rules, have been widely studied. The purpose of this work is to present a detailed study of the natural optimization method called Differential Evolution and its strategies. A theoretical formulation is presented. In this work, a revision of Genetic Algorithms and Simulated Annealing is made. These techniques are employed to compare their results to those obtained with Differential Evolution. Classic mathematical functions and some problems of engineering are used in order to verify the efficiency of the studied technique. The Differential Evolution is applied with effectiveness in multi-objective optimization problems with and without the presence of constraints including two complex robotic problems. / Nas últimas décadas, a aplicação de otimização em problemas de engenharia tem crescido consideravelmente. Existem muitos métodos de otimização e o desempenho de cada um deles depende do tipo de problema considerado. Os métodos naturais, que se baseiam em regras probabilísticas, têm sido amplamente estudados. O objetivo deste trabalho é apresentar um estudo detalhado do método de otimização natural denominado Evolução Diferencial e suas estratégias, apresentando sua fundamentação teórica. Neste trabalho é realizada uma revisão de Algoritmos Genéticos e Recozimento Simulado. Estas técnicas são utilizadas para comparar os resultados por elas obtidos com os calculados aplicando Evolução Diferencial. Para verificar a eficiência do método estudado, são utilizados funções matemáticas clássicas e alguns problemas de engenharia. A Evolução Diferencial é aplicada com eficiência em problemas de otimização multi-objetivo, na presença ou não de restrições, incluindo dois problemas complexos em robótica. / Mestre em Engenharia Mecânica

O presente trabalho otimizou os parâmetros concentrados do sistema de suspensão de um modelo de veículo completo, representando um automóvel de passeio que trafega a uma velocidade constante por um determinado perfil de pista previsto na norma ISO 8608, 1995, através da utilização de um algoritmo meta-heurístico de otimização multiobjetivo. Duas rotinas numérico-computacionais foram desenvolvidas, visando realizar tal otimização tanto no domínio do tempo quanto no domínio da frequência. A utilização de algoritmos meta-heurísticos vem ganhando espaço na otimização de sistemas mecânicos, proporcionando rapidez e precisão na obtenção de resultados ótimos. Ao se combinar um algoritmo de otimização a um modelo que represente satisfatoriamente um sistema mecânico, obtém-se uma ferramenta indicadora dos parâmetros de máxima eficiência do sistema, que pode ser utilizada em inúmeras aplicações. Pretendeu-se, com a integração de rotinas de análise dinâmica nos domínios do tempo e da frequência ao algoritmo genético de otimização multiobjetivo NSGA-II, desenvolvido por Deb et al., 2002, a obtenção de duas fronteiras ótimas de Pareto. Estas fronteiras consistem no conjunto de soluções não dominadas que minimizam as seguintes funções objetivo: o valor RMS ponderado da aceleração vertical do assento do motorista, o valor RMS da média do fator de amplificação dinâmica das quatro rodas do modelo e o máximo deslocamento relativo entre cada roda e a carroceria. O método proposto por Shinozuka e Jan, 1972, é utilizado para a obtenção do perfil de irregularidades da pista no domínio do tempo a partir das equações de densidade espectral de potência (PSD) que representam as diferentes classes de pavimentos. O método de Newmark, 1959, é utilizado para resolver a equação diferencial de movimento no domínio do tempo e obter a resposta dinâmica do modelo a tais irregularidades. O comportamento dinâmico do modelo de veículo no domínio da frequência foi obtido através da utilização da função de resposta em frequência (FRF) do modelo de veículo analisado. Os resultados demonstraram a capacidade de ambas as rotinas de análise dinâmica desenvolvidas de produzir resultados consistentes com os encontrados na literatura, bem como a capacidade dos algoritmos de otimização implementados de fornecer fronteiras ótimas de Pareto para os problemas propostos. / The proposed work optimized the concentrated parameters of a full-vehicle model’s suspension system, being that model representative of a passenger car which travels at a constant speed on a certain road profile provided by the ISO 8608, 1995, standard, using a multi-objective meta-heuristic optimization algorithm. Two numerical-computational routines were developed, seeking to perform said optimization for both the time and frequency domains. The use of meta-heuristic algorithms has been increasing in mechanical systems optimization, providing speed and accuracy in obtaining an optimal result. Combining an optimization algorithm with a model that satisfactorily represents a mechanical system yields a tool that indicates the system’s maximum efficiency parameters, which can be used in numerous applications. It was intended, with the integration of the dynamic analysis routines to the multi-objective genetic optimization algorithm NSGA-II, developed by Deb et al., 2002, the obtainment of two Pareto-optimal fronts. These fronts consist in the set of non-dominated solutions that minimize the following objective functions: the weighted RMS value of the driver’s seat vertical acceleration, the mean RMS value of the model wheel’s dynamic amplification factor, and the maximum relative displacement between each wheel and the body of the vehicle model. The method proposed by Shinozuka and Jan, 1972, is used to obtain the road irregularity profile in the time domain from the power spectral density (PSD) equations that represent the different pavement classes. The Newmark’s method (1959) is used to solve the differential motion equation in the time domain, in order to obtain the vehicle model’s responses to these irregularities. The dynamic behavior of the vehicle model in the frequency domain was obtained through the use of the frequency response function (FRF) of the analyzed model. The results showed the capacity of both the dynamic analysis routines developed in generating results that are consistent with those found in literature, as well as the capacity of the optimization algorithms implemented in providing Pareto optimal fronts to the proposed problems.

The problem of designing control systems for entire plants is studied. A review of previous works, available techniques and current research challenges is presented, followed by the description of some theoretical tools to improve plantwide control, including the proposal of an augmented lexicographic multi-objective optimization procedure. With these, an augmented hierarchical plantwide control design technique and an optimal multi-objective technique for integrated control structure selection and controller tuning are proposed. The main contributions of these proposed techniques are the inclusion of system identification and optimal control tuning as part of the plantwide design procedure for improved results, support to multi-objective control specifications and support to any type of plant and controllers. Finally, the proposed techniques are applied to industrial benchmarks to demonstrate and validate its applicability. / O problema de projetar sistemas de controle para plantas inteiras é estudado. Uma revisão de trabalhos anteriores, técnicas disponíveis e atuais desafios de pesquisa é apresentada, seguida da descrição de algumas ferramentas teóricas para melhorar o controle plantwide, incluindo a proposta de um procedimento de otimização multi-objetivo lexicográfico aumentado. Com tais elementos, são propostas uma nova técnica hierárquica aumentada de projeto de sistemas de controle plantwide e uma técnica multi-objetivo para seleção de estrutura de controlador integrada à sintonia ótima do controlador. As principais contribuições das técnicas propostas são a inclusão de identificação de sistemas e sintonia ótima de controladores como parte do procedimento de projeto de controle plantwide para melhores resultados, suporte a especificações multi-objetivo e suporte a quaisquer tipos de plantas e controladores. Finalmente, as técnicas propostas são aplicadas a benchmarks industriais para demonstrar e validar sua aplicabilidade.