Dans ce travail, une méthode d'optimisation à coût fini, essentiellement locale,mais qui devient globale lorsque le nombre d'analyses croît est développée. La ± globalisation α vient de ré-initialisations probabilisées de recherches locales prenant en compte les points de départ et de convergence passés. L'optimiseur local est une version améliorée de la méthode de Nelder-Mead, où les variables sont bornées, où les contraintes d'inégalité sont prises en compte par pénalisation adaptative, et où les dégénérescences du simplexe sont traitées par ré-initialisation. Cette méthode, appelée ± Globalized and Bounded Nelder-Mead α (GBNM), est testée sur des fonctions multimodales et des problèmes de conception de stratifiés composites. Puis, des applications plus complexes sont traitées avec GBNM : l'optimisation de la raideur de flexion et l'identification des positions des sauts de plis de monopalmes de nage.

Before a new flight control system is released for flight, a huge number of simulations are evaluated to find weaknesses of the system. This process is called flight clearance. Flight clearance is a very important but time consuming process. There is a need of better flight clearance methods and one of the most promising methods is the use of optimization. In this thesis the flight clearance of a simulation model of JAS 39 Gripen is examined. Two flight clearance algorithms using two different optimization methods are evaluated and compared to each other and to a traditional flight clearance method. In this thesis the flight clearance process is separated into three cases: search for the worst flight condition, search for the worst manoeuvre and search for the worst flight condition including parameter uncertainties. For all cases the optimization algorithms find a more dangerous case than the traditional method. In the search for worst flight condition, both with and without uncertainties, the optimization algorithms are to prefer to the traditional method with respect to the clearance results and the number of objective function calls. The search for the worst manoeuvre is a much more complex problem. Even as the algorithms find more dangerous manoeuvres than the traditional method, it is not certain that they find the worst manoeuvres. If not other methods should be used the problem has to be rephrased. For example other optimization variables or a few linearizations of the optimization problem could reduce the complexity. The overall impression is that the need of information and problem characteristics define which method that is most suitable to use. The information required must be weighed against the cost of objective function calls. Compared to the traditional method, the optimization methods used in this thesis give extended information about the problems examined and are better to locate the worst case.

No description available.

Ce mémoire présente les travaux réalisés en aérodynamique afin de pouvoir disposer d’une chaîne de conception complète nécessaire au traitement des différentes problématiques sur les pales éoliennes, qui ne peuvent être pris en compte à l’aide d’une méthode unique. Afin de pouvoir faire de l’optimisation de pales en fonction de différents critères, un outil numérique simplifié (VALDAG) a été développé. Le module de simulation utilise la méthode du Disque Actif Généralisé, qui s’appuie sur la résolution des équations Navier-Stokes, complété par des corrections empiriques. Il respecte un compromis entre la précision et le coût de calcul. Cet outil pour lequel une interface web a été développée pour le rendre adapté à des travaux d’ingénierie est susceptible de se calibrer automatiquement sur une nouvelle géométrie de pale grâce à la méthode Nelder-Mead Simplex. On peut ensuite réaliser une optimisation de la performance de la pale en modifiant cette géométrie, et générer les fichiers 3D qui serviront d’entrée à la simulation 3D et de visualisation des optimisations de forme. Ces designs optimisés sont ensuite validés par des simulations DNS à l’aide de l’outil NaSCar 3D développé à l’INRIA. Cet outil résout des équations Navier-Stokes sur un maillage cartésien 3D et prend en compte des obstacles immergés via la fonction Level-Set et la pénalisation. Après des adaptations sur le traitement de la géométrie de pales, un compromis de configuration CFD est trouvé pour simuler un rotor éolien. En conclusion, ce paquet d’outils VALDAG est peu coûteux, facile à utiliser et efficace. En associant VALDAG aux simulations 3D, une chaîne de conception est complétée. / This Ph.D. thesis presents some research work on aerodynamics of wind turbine blades, in order to dispose a conception chain necessary for different problems, which cannot be treated by one unique method. A simplified numerical toolkit (VALDAG) has been developed to optimize the performance of blades in different creteria. The simulation module use the Generalized Actuator Disc model, which relies on the solution of Navier-Stokes equations and completed with empiric corrections. This tool respects a reasonable compromise between model complexity and computational reliability. An automatic calibration mechanism was implemented using the Nelder-Mead Simplex algorithm. A web users interface (WUI) is also available to adapt VALDAG for industrial engineers. Optimization is then carried on by modifying the blades’ geometry parameters and the designs optimized is stocked in files which can be used for 3D simulation and/or visualization. The blade designed with VALDAG are then simulated by a 3D numerical tool to validate previous predictions. This 3D tool called NaSCar 3D is developed in INRIA and resolve Navier-Stokes equations on to a cartesian mesh, in which the immersed obstacles are considered with the Level-Set function and the penalization method. After some necessary adaptation for the treatment of blades’ geometry, a compromise on CFD configuration is found to simulate a wind turbine rotor. To conclude, this design toolkit VALDAG is low time-costly, user-friendly and efficient. Associated with 3D simulations, a conception chain has been established.

Made available in DSpace on 2014-07-29T15:08:22Z (GMT). No. of bitstreams: 1 Adson.pdf: 1251737 bytes, checksum: 08b3122009f15a40cec4f32fda3231e9 (MD5) Previous issue date: 2008-11-07 / The present work deals with the problem of planning the distribution system of electricity and is divided into three parts: Problem Definition, Resolution Approaches and Results and Conclusions. The energy distribution networks are of great economic importance in countries like Brazil. On one hand, there are fixed costs of physical installation and operation of the network, mainly due to the costs of energy losses and, secondly, the natural obstacles along the possible passages of network s links. The large amount of these costs, together with lack of efficient methods when it comes to real applications in the matter, justify the development of this research. The study of such aspects, the precise definition of the problem and the reasons that motivated this work can be found on the first part of this work. The second part shows the approaches for resolution. Three proposals methods were adopted: the first uses the algorithm Prim associated with the method Nelder-Mead Simplex. In the second proposal uses Dynamic Programming and, finally, we take the metaphor of Ant Colony also associated with the Nelder-Mead Simplex. The results, presented at the third part of this work, demonstrated the effectiveness of the proposed methods, especially the good compromise between performance and applicability obtained by the third proposal. / O presente trabalho lida com o problema de planejamento da rede de distribuição de energia elétrica, estando dividido em três partes: Definição do Problema, Abordagens de Resolução e Resultados e Conclusões. As redes de distribuição de energia têm uma grande importância econômica em países como o Brasil. Por um lado, há os custos físicos fixos de instalação e de operação da rede, sobretudo os custos devido às perdas de energia e, por outro, os obstáculos naturais impostos ao longo das possíveis passagens para as ligações da rede. O montante elevado destes custos, unidos à escassez de métodos eficientes quando se trata de aplicações reais no assunto, justificam o desenvolvimento desta pesquisa. O estudo de tais aspectos, a definição precisa do problema e as justificativas podem ser encontradas na primeira parte deste trabalho. Na segunda parte, apresentam-se as abordagens de resolução. Três propostas foram adotadas: na primeira usa-se o algoritmo Prim associado ao método Nelder-Mead Simplex. Na segunda proposta utiliza-se a Programação Dinâmica e, por fim, tomamos a metáfora de colônia de formigas também associada ao Nelder-Mead Simplex. Os resultados, apresentados na terceira parte deste trabalho, mostram a eficácia dos métodos propostos, em especial o bom compromisso entre performance e aplicabilidade obtido pela terceira proposta.

This master's thesis deals with the optimization of numerical coefficients of the Wiener filter for muscle noise filtering in ECG signals. The theoretical part deals with ECG signal characteristic and muscle interference. It also contains a summary of the wavelet transform, wavelet Wiener's filtration, methods for calculating of the threshold and thresholding. In the last theoretical part the characteristic optimization techniques, the exhausive search and Nelder-Mead simplex method are mentioned, which were implemented in the practical part of this thesis in MATLAB. The functional verification and Wiener's filter optimization were tested on the standard electrocardiograms database CSE. By using the methods of exhausive search, the initial estimate for the solution method Nelder-Mead was obtained. The optimization method Nelder-Mead gives better results in the orders of hundredths or tenths than the method of exhausive search. The practical part is finished by the comparison of results of implemented algorithm with optimum coefficients, implemented in this thesis, with the results of other methods for filtering muscle interference in ECG signals.

Orientador: Edevar Luvizotto Junior / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Civil, Arquitetura e Urbanismo / Made available in DSpace on 2018-08-05T06:54:25Z (GMT). No. of bitstreams: 1 Gumier_CarlosCesar_M.pdf: 3470729 bytes, checksum: 2e93f4938a5eb678d7fb52827439bbc5 (MD5) Previous issue date: 2005 / Resumo: Este trabalho apresenta, sob diversos aspectos, as perdas de água que ocorrem nas redes de distribuição dos sistemas de abastecimento e os mecanismos de controle disponíveis. Dentre os objetivos se destaca a proposição do modelo matemático-computacional como uma nova ferramenta de localização de fugas. O modelo proposto foi desenvolvido através do acoplamento de um simulador hidráulico baseado no Time Marching Approach ¿ TMA com o algoritmo otimizador de Nelder-Mead. O trabalho visa ampliar os resultados obtidos nas investigações desenvolvidas em LUVIZOTTO (1998) e OCAMPOS (2003), aplicando o modelo em uma rede de distribuição de água real, utilizando os recursos disponibilizados pela companhia de água, propondo uma metodo de aplicação do modelo para tais condições / Abstract: This work presents, under diverse aspects, the water losses that occur in the distribution networks of the water supply systems and the available mechanisms of control. Amongst the objectives, it detaches the proposal of a computational model as a new tool for leak localization. The considered model was developed through the coupling of hydraulic simulator based in Time Marching Approach ¿ TMA method with the Nelder-Mead optimization algorithm. The work aims to extend the results obtained by LUVIZOTTO (1998) and OCAMPOS (2003), by applying the model to a real water distribution network, using the data provided by the water company and suggesting a methodology for application of the model for such conditions / Mestrado / Recursos Hidricos / Mestre em Engenharia Civil

Derivatives have a large and significant role on the financial markets today and the popularity of options has increased. This has also increased the demand of finding a suitable option pricing model, since the ground-breaking model developed by Black & Scholes (1973) have poor pricing performance. Practitioners and academics have over the years developed different models with the assumption of non-constant volatility, without reaching any conclusions regarding which model is more suitable to use. This thesis examines four different models, the first model is the Practitioners Black & Scholes model proposed by Christoffersen & Jacobs (2004b). The second model is the Heston´s (1993) continuous time stochastic volatility model, a modification of the model is also included, which is called the Strike Vector Computation suggested by Kilin (2011). The last model is the Heston & Nandi (2000) Generalized Autoregressive Conditional Heteroscedasticity type discrete model. From a practical point of view the models are evaluated, with the goal of finding the model with the best pricing performance and the most practical usage. The model´s robustness is also tested to see how the models perform in out-of-sample during a high respectively low implied volatility market. All the models are effected in the robustness test, the out-sample ability is negatively affected by a high implied volatility market. The results show that both of the stochastic volatility models have superior performances in the in-sample and out-sample analysis. The Generalized Autoregressive Conditional Heteroscedasticity type discrete model shows surprisingly poor results both in the in-sample and out-sample analysis. The results indicate that option data should be used instead of historical return data to estimate the model’s parameters. This thesis also provides an insight on why overnight-index-swap (OIS) rates should be used instead of LIBOR rates as a proxy for the risk-free rate.

Forests have been proposed as a means of reducing atmospheric carbon dioxide levels due to their ability to store carbon as biomass. To quantify the amount of atmospheric carbon sequestered by forests, biomass and density estimates are often needed. This study develops, implements, and tests an individual tree-based algorithm for obtaining forest density and biomass using orthophotographs and small footprint LiDAR imagery. It was designed to work with a range of forests and image types without modification, which is accomplished by using generic properties of trees found in many types of images. Multiple parameters are employed to determine how these generic properties are used. To set these parameters, training data is used in conjunction with an optimization algorithm (a modified Nelder-Mead simplex algorithm or a genetic algorithm). The training data consist of small images in which density and biomass are known. A first test of this technique was performed using 25 circular plots (radius = 15 m) placed in young pine plantations in central Virginia, together with false color othophotograph (spatial resolution = 0.5 m) or small footprint LiDAR (interpolated to 0.5 m) imagery. The highest density prediction accuracies (r2 up to 0.88, RMSE as low as 83 trees / ha) were found for runs where photointerpreted densities were used for training and testing. For tests run using density measurements made on the ground, accuracies were consistency higher for orthophotograph-based results than for LiDAR-based results, and were higher for trees with DBH ≥10cm than for trees with DBH ≥7 cm. Biomass estimates obtained by the algorithm using LiDAR imagery had a lower RMSE (as low as 15.6 t / ha) than most comparable studies. The correlations between the actual and predicted values (r2 up to 0.64) were lower than comparable studies, but were generally highly significant (p ≤ 0.05 or 0.01). In all runs there was no obvious relationship between accuracy and the amount of training data used, but the algorithm was sensitive to which training and testing data were selected. Methods were evaluated for combining predictions made using different parameter sets obtained after training using identical data. It was found that averaging the predictions produced improved results. After training using density estimates from the human photointerpreter, 89% of the trees located by the algorithm corresponded to trees found by the human photointerpreter. A comparison of the two optimization techniques found them to be comparable in speed and effectiveness. / Ph. D.

En ce début de 21e siècle, le bruit des transports est l'une des principales nuisances. Un important travail est réalisé par les acousticiens afin d'améliorer l'efficacité des écrans acoustiques.<br />Actuellement, le choix des protections anti-bruit se porte souvent sur des solutions triviales ou issues d'études paramétriques. Or le CSTB possède des codes de calculs numériques avancés dédiés à la simulation de la propagation acoustique en milieu extérieur. Le but de cette thèse est de coupler ces codes prévisionnels avec un outil mathématique à développer permettant une recherche systématique des solutions optimales.<br />Une étude bibliographique a permis de dégager des algorithmes applicables au cas de l'optimisation de paramètres caractérisant un écran acoustique. Ils ont été appliqués avec succès aux écrans multi-diffracteurs. Des méthodes d'optimisation multiobjectif ont en outre permis d'optimiser simultanément le prix de revient de la protection et son efficacité acoustique. Dans une optique de développement durable, un travail a été réalisé sur l'efficacité de protections acoustiques de faible hauteur en milieu urbain dans le but de créer des espaces calmes. Des mesures sur modèles réduits ont été confrontées aux simulations acoustiques et l'optimisation de la forme de ces protections a permis de dégager des solutions optimales. Enfin, les méthodes d'optimisation ont été appliquées au cas de l'optimisation de trajectoires.<br />Les principaux objectifs de la thèse ont été atteints, notamment par la création d'un outil utilisant des algorithmes d'optimisation (Nelder Mead et Algorithmes Génétiques) ainsi que les codes de calculs de simulation acoustique en milieu extérieur.