Accurate transistor models are important in wireless and microwave circuit design. Large-signal field-effect transistor (FET) models are generally extracted from current-voltage (IV) characteristics, small-signal S-parameters, and large-signal measurements. This dissertation describes improved characterization and measurement validation techniques for FET models that correctly account for thermal and trapping effects. Demonstration of a customized pulsed-bias, pulsed-RF S-parameter system constructed by the author using a traditional vector network analyzer is presented, along with the design of special bias tees to allow pulsing of the bias voltages. Pulsed IV and pulsed-bias S-parameter measurements can provide results that are electrodynamically accurate; that is, thermal and trapping effects in the measurements are similar to those of radio-frequency or microwave operation at a desired quiescent bias point. The custom pulsed S-parameter system is benchmarked using passive devices and advantages and tradeoffs of pulsed S-parameter measurements are explored. Pulsed- and continuous-bias measurement results for a high-power transistor are used to validate thermal S-parameter correction procedures. A new implementation of the steepest-ascent search algorithm for load-pull is presented. This algorithm provides for high-resolution determination of the maximum power and associated load impedance using a small number of measured or simulated reflection-coefficient states. To perform a more thorough nonlinear model validation, it is often desired to find the impedance providing maximum output power or efficiency over variations of a parameter such as drain voltage, input power, or process variation. The new algorithm enables this type of validation that is otherwise extremely tedious or impractical with traditional load-pull. A modified nonlinear FET model is presented in this work that allows characterization of both thermal and trapping effects. New parameters and equation terms providing a trapping-related quiescent-bias dependence have been added to a popular nonlinear ("Angelov") model. A systematic method for fitting the quiescent-dependence parameters, temperature coefficients, and thermal resistance is presented, using a GaN high electron-mobility transistor as an example. The thermal resistance providing a good fit in the modeling procedure is shown to correspond well with infrared measurement results.

Submitted by Geandra Rodrigues (geandrar@gmail.com) on 2018-03-27T14:05:45Z No. of bitstreams: 1 franciscocarlosrodriguescoelho.pdf: 4772391 bytes, checksum: e11633134429c05832808dad96be9940 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2018-03-27T14:28:31Z (GMT) No. of bitstreams: 1 franciscocarlosrodriguescoelho.pdf: 4772391 bytes, checksum: e11633134429c05832808dad96be9940 (MD5) / Made available in DSpace on 2018-03-27T14:28:31Z (GMT). No. of bitstreams: 1 franciscocarlosrodriguescoelho.pdf: 4772391 bytes, checksum: e11633134429c05832808dad96be9940 (MD5) Previous issue date: 2018-02-22 / A alocação de Geração Distribuída (GD) em sistemas de distribuição de energia elétrica consiste em determinar os barramentos para conexão destas unidades geradoras, e o montante de potência a ser injetado, visando um ou mais objetivos, que podem ser: redução das perdas de potência ativa, melhorias no perfil de tensão, minimização dos custos operacionais, maximização da geração de energia, ganhos ambientais, dentre outros. O principal objetivo considerado neste trabalho é a minimização das perdas de potência ativa, mantendo as tensões dos barramentos dentro de limites recomendados. Para alcançar este objetivo, uma metodologia de otimização é proposta, tratando separadamente os problemas de localização das unidades geradoras no sistema, e o dimensionamento destas unidades. A determinação das barras com conexão de GD é realizada através de uma nova técnica de otimização metaheurística, implementada no MATLAB, denominada Metaheurística Empírica Discreta (MED). Já o dimensionamento das unidades de GD é realizado de duas formas distintas, a depender do tipo de sistema de distribuição analisado. No caso dos sistemas cujos dados são equivalentes monofásicos, o montante de potencia é determinado por um Fluxo de Potência Ótimo implementado no software comercial LINGO. A segunda estratégia de determinação da potência despachada é empregada no caso dos testes realizados com sistemas trifásicos desbalanceados, cujo dimensionamento é feito pelo método do gradiente descendente e o cálculo do fluxo de potência é realizado pelo software OpenDSS. Os três sistemas equivalentes monofásicos utilizados são compostos por 33, 69 e 476 barras, enquanto os dois trifásicos desequilibrados possuem 34 e 123 barras. A qualidade da metodologia proposta na resolução do problema de alocação de geração distribuída é avaliada através de comparações com a literatura especializada, comparações com outras metaheurísticas e testes de robustez. Os resultados provenientes de simulações com alocação de três e quatro unidades de GD em sistemas de distribuição de energia elétrica mostram que a metodologia proposta é eficiente, sendo capaz de produzir resultados com significativas reduções nas perdas de potência ativa e perfis de tensão adequados. / The optimal Distributed Generation (DG) allocation problem consists in choosing the best locations of those distributed power plants at the distribution system, and to define its amount of power injection. The approach can be either single or multiobjective. The main objectives are: minimization of total power loss, voltage profile improvement, operational cost minimization, maximization of distributed generation capacity, environmental gains, among others. In this work, the main goal pursued is the total power loss minimization of the distribution system, keeping the buses voltages within the predetermined limits. To achieve this goal, an optimization methodology is proposed. This approach treats separately the location problem and the power dispatched by the generation units. The busbars connected to distributed generation are determined through a new metaheuristic algorithm, implemented in MATLAB, named Empirical Discrete Metaheuristic (EDM). The amount of power injection is solved by an Optimum Power Flow implemented in the commercial software LINGO, or by the Steepest Descent Method in the MATLAB environment. The first strategy to determine the DG dispatch is used on simulations with single phase equivalents systems. The second one is employed in the amount of power determination in unbalanced three phase systems, which the power flow is carried out by the open source software OpenDSS. The three single phase equivalent test systems analyzed are composed by 33, 69 and 476 buses, while the two systems with three phases have 34 and 123 buses, each. To evaluate the proposed methodology quality, comparisons to published works in the specialized literature are made. Also, robustness tests and comparisons to other well succeed metaheuristics are carried out. The results were obtained from simulations with three and four DG units in electric power distribution systems. These results consistently show that the proposed methodology is efficient, providing DGs configurations that significantly reduces the active power losses and keep the voltages at adequate levels.

The aim of this master thesis is to study algorithms for automatically tuning antenna parameters to improve the performance of the radio access part of a telecommunication network and user experience. There are four dierent optimization algorithms, Stepwise Minimization Algorithm, Random Search Algorithm, Modied Steepest Descent Algorithm and Multi-Objective Genetic Algorithm to be applied to a model of a radio access network. The performances of all algorithms will be evaluated in this thesis. Moreover, a graphical user interface which is developed to facilitate the antenna tuning simulations will also be presented in the appendix of the report.

No description available.

O objetivo principal deste trabalho é o estudo da regularidade anallítica global de certos operadores diferenciais definidos no toro. Uma ferramenta fundamental utilizada neste estudo são as séries parciais de Fourier, que nos permitem caracterizar tanto as distribuições periódicas quanto as funções anallíticas reais periódicas através do comportamento assintótico de seus coeficientes parciais de Fourier. Neste sentido, apresentamos também um estudo detalhado das relações destes objetos com seus coeficientes parciais de Fourier / The main goal of this work is to study global analytic regularity properties of certain differential operators acting in the torus. A main tool that will be used to achieve our goals are the partial Fourier series, which allow us to characterize objects such as periodic distributions or periodic real analytic functions in terms of the growth of their partial Fourier coefficients

O objetivo principal deste trabalho é o estudo da regularidade anallítica global de certos operadores diferenciais definidos no toro. Uma ferramenta fundamental utilizada neste estudo são as séries parciais de Fourier, que nos permitem caracterizar tanto as distribuições periódicas quanto as funções anallíticas reais periódicas através do comportamento assintótico de seus coeficientes parciais de Fourier. Neste sentido, apresentamos também um estudo detalhado das relações destes objetos com seus coeficientes parciais de Fourier / The main goal of this work is to study global analytic regularity properties of certain differential operators acting in the torus. A main tool that will be used to achieve our goals are the partial Fourier series, which allow us to characterize objects such as periodic distributions or periodic real analytic functions in terms of the growth of their partial Fourier coefficients

Concorder un modèle à certaines observations, voilà qui résume assez bien ce que l’apprentissage machine cherche à accomplir. Ce concept est maintenant omniprésent dans nos vies, entre autre grâce aux percées récentes en apprentissage profond. La stratégie d’optimisation prédominante pour ces deux domaines est la minimisation d’un objectif donné. Et pour cela, la méthode du gradient, méthode de premier-ordre qui modifie les paramètres du modèle à chaque itération, est l’approche dominante. À l’opposé, les méthodes dites de second ordre n’ont jamais réussi à s’imposer en apprentissage profond. Pourtant, elles offrent des avantages reconnus qui soulèvent encore un grand intérêt. D’où l’importance de la méthode du col, qui unifie les méthodes de premier et second ordre sous un même paradigme. Dans ce mémoire, nous établissons un parralèle direct entre la méthode du col et le domaine du contrôle optimal ; domaine qui cherche à optimiser mathématiquement une séquence de décisions. Et certains des problèmes les mieux compris et étudiés en contrôle optimal sont les commandes linéaires quadratiques. Problèmes pour lesquels on connaît très bien la solution optimale. Plus spécifiquement, nous démontrerons l’équivalence entre une itération de la méthode du col et la résolution d’une Commande Linéaire Quadratique (CLQ). Cet éclairage nouveau implique une approche unifiée quand vient le temps de déployer nombre d’algorithmes issus de la méthode du col, tel que la méthode du gradient et celle des gradients naturels, sans être limitée à ceux-ci. Approche que nous étendons ensuite aux problèmes à horizon infini, tel que les modèles à équilibre profond. Ce faisant, nous démontrons pour ces problèmes que calculer les gradients via la différentiation implicite revient à employer l’équation de Riccati pour solutionner la CLQ associée à la méthode du gradient. Finalement, notons que l’incorporation d’information sur la courbure du problème revient généralement à rencontrer une inversion matricielle dans la méthode du col. Nous montrons que l’équivalence avec les CLQ permet de contourner cette inversion en utilisant une approximation issue des séries de Neumann. Surprenamment, certaines observations empiriques suggèrent que cette approximation aide aussi à stabiliser le processus d’optimisation quand des méthodes de second-ordre sont impliquées ; en agissant comme un régularisateur adaptif implicite. / Machine learning entails training a model to fit some given observations, and recent advances in the field, particularly in deep learning, have made it omnipresent in our lives. Fitting a model usually requires the minimization of a given objective. When it comes to deep learning, first-order methods like gradient descent have become a default tool for optimization in deep learning. On the other hand, second-order methods did not see widespread use in deep learning. Yet, they hold many promises and are still a very active field of research. An important perspective into both methods is steepest descent, which allows you to encompass first and second-order approaches into the same framework. In this thesis, we establish an explicit connection between steepest descent and optimal control, a field that tries to optimize sequential decision-making processes. Core to it is the family of problems known as Linear Quadratic Regulation; problems that have been well studied and for which we know optimal solutions. More specifically, we show that performing one iteration of steepest descent is equivalent to solving a Linear Quadratic Regulator (LQR). This perspective gives us a convenient and unified framework for deploying a wide range of steepest descent algorithms, such as gradient descent and natural gradient descent, but certainly not limited to. This framework can also be extended to problems with an infinite horizon, such as deep equilibrium models. Doing so reveals that retrieving the gradient via implicit differentiation is equivalent to recovering it via Riccati’s solution to the LQR associated with gradient descent. Finally, incorporating curvature information into steepest descent usually takes the form of a matrix inversion. However, casting a steepest descent step as a LQR also hints toward a trick that allows to sidestep this inversion, by leveraging Neumann’s series approximation. Empirical observations provide evidence that this approximation actually helps to stabilize the training process, by acting as an adaptive damping parameter.

This Master Thesis deals with image restoration using deconvolution. The terms introducing into deconvolution theory like two-dimensional signal, distortion model, noise and convolution are explained in the first part of thesis. The second part deals with deconvolution methods via utilization of the Bayes approach which is based on the probability principle. The third part is focused on the Alternating Minimization Algorithm for Multichannel Blind Deconvolution. At the end this algorithm is written in Matlab with utilization of the NAG C Library. Then comparison of different optimization methods follows (simplex, steepest descent, quasi-Newton), regularization forms (Tichonov, Total Variation) and other parameters used by this deconvolution algorithm.