Dois modelos de controle de risco: o modelo Nelson-Siegel dinâmico e o cálculo de VaR por modelos GARCH

A presente dissertação tem como objetivo apresentar dois importantes modelos usados na análise de risco. Essa análise culmina em uma aplicação empírica para cada um deles. Apresenta-se primeiro o modelo Nelson-Siegel dinâmico, que estima a curva de juros usando um modelo paramétrico exponencial parcimonioso. É citada a referência criadora dessa abordagem, que é Nelson & Siegel (1987), passa-se pela apresentação da mais importante abordagem moderna que é a de Diebold & Li (2006), que é quem cria a abordagem dinâmica do modelo Nelson-Siegel, e que é inspiradora de diversas extensões. Muitas dessas extensões também são apresentadas aqui. Na parte empírica, usando dados da taxa a termo americana de Janeiro de 2004 a Março de 2015, estimam-se os modelos Nelson-Siegel dinâmico e de Svensson e comparam-se os resultados numa janela móvel de 12 meses e comparamos seus desempenhos com aqueles de um passeio aleatório. Em seguida, são apresentados os modelos ARCH e GARCH, citando as obras originais de Engle (1982) e Bolleslev (1986) respectivamente, discutem-se características destes modelos e apresentam-se algumas extensões ao modelo GARCH, incluindo aí alguns modelos GARCH multivariados. Passa-se então por uma rápida apresentação do conceito de VaR (Value at Risk), que será o objetivo da parte empírica. Nesta, usando dados de 02 de Janeiro de 2004 até 25 de Fevereiro de 2015, são feitas uma estimação da variância de um portfólio usando os modelos GARCH, GJR-GARCH e EGARCH e uma previsão do VaR do portfólio a partir da estimação feita anteriormente. Por fim, são apresentados alguns trabalhos que usam os dois modelos conjuntamente, ou seja, que consideram que as taxas ou os fatores que as podem explicam possuem variância variante no tempo.

Combinação social e construção de equipes: um estudo sobre os fatores relevantes para o processo de composição de equipes na educação executiva brasileira

A composição de equipes é um tema recorrente em diferentes áreas do conhecimento. O interesse pela definição das etapas e variáveis relevantes desse processo, considerado complexo, é manifestado por pesquisadores, profissionais e desenvolvedores de Sistemas de Informação (SI). Todavia, enquanto linhas teóricas, oriundas dos estudos organizacionais, buscam a consolidação de modelos matemáticos que reflitam a relação entre variáveis de composição de equipes e o seu desempenho, teorias emergentes, como a de Combinação Social, acrescentam novos elementos à discussão. Adicionalmente, variáveis específicas de cada contexto, que no caso dessa pesquisa é a educação executiva brasileira, também são mencionadas como tendo relevância para estruturação de grupos. Dado o interesse e a variedade de vertentes teóricas que abordam esse fenômeno, essa pesquisa foi proposta para descrever como ocorre a construção de equipes docentes e identificar as variáveis consideradas relevantes neste processo. Um modelo teórico inicial foi desenvolvido e aplicado. Dada a característica da questão de pesquisa, foi utilizada uma abordagem metodológica exploratório-descritiva, baseada em estudos de casos múltiplos, realizados em quatro instituições de ensino superior brasileiras, que oferecem cursos de educação executiva. A coleta e a análise de dados foi norteada pelos métodos propostos por Huberman e Miles (1983) e Yin (2010), compreendendo a utilização de um protocolo de estudo de caso, bem como o uso de tabelas e quadros, padronizados à luz do modelo teórico inicial. Os resultados desse trabalho indicam, majoritariamente, que: as teorias de Combinação Social e as teorias de Educação adicionam elementos que são relevantes ao entendimento do processo de composição de equipes; há variáveis não estruturadas que deixam de ser consideradas em documentos utilizados na avaliação e seleção de profissionais para equipes docentes; e há variáveis de composição que só são consideradas após o fim do primeiro ciclo de atividades das equipes. Com base nos achados empíricos, a aplicação do modelo teórico foi ajustada e apresentada. As contribuições adicionais, as reflexões, as limitações e as propostas de estudos futuros são apresentadas no capítulo de conclusões.

Simulação do comportamento espectral de redes de período longo em fibra óptica

Esta tese é dedicada às redes de período longo, LPG’s, em fibra óptica, escritas por exposição a radiação ultravioleta (UV) ou por exposição a descargas eléctricas, e às suas aplicações em comunicações ópticas e em sistemas sensores. Numa primeira fase estudaram-se os aspectos teóricos fundamentais para a compreensão das LPG, nomeadamente os dois modelos matemáticos propostos na literatura, para a transmissão espectral de uma LPG, o modelo de duas camadas e o modelo de três camadas. Em seguida, estudou-se o deslocamento do comprimento de onda ressonante perante mudanças de parâmetros externos. Aqui, verificou-se que para variações da temperatura no exterior da LPG, o deslocamento do comprimento de onda ressonante é linear. Por outro lado, para variações de índice de refracção exterior, verificou-se que à medida que o índice exterior se aproxima dos valores do índice de refracção da bainha, o comprimento de onda ressonante tende para valores mais baixos. Por último, realizou-se um estudo da transmissão espectral de duas aplicações que envolvem LPG’s, nomeadamente dois tipos de interferómetros e filtros ópticos. Numa segunda fase, foi desenvolvida uma ferramenta de simulação destes modelos, que permitia não só a obtenção dos espectros de transmissão das LPG’s mas também a obtenção das curvas de phase matching em função do período da rede e do comprimento de onda ressonante. A aplicação permitia também a obtenção das curvas do deslocamento do comprimento de onda ressonante, perante variações do índice de refracção exterior ou da temperatura. Para além disso, essa ferramenta realiza a simulação dos espectros de transmissão de filtros ópticos e de interferómetros de Michelson e de Mach-Zehnder construídos com base em LPG’s. A última fase do trabalho, a componente laboratorial, foi realizada na Unidade de Optoelectrónica e Sistemas Electrónicos do INESC Porto, onde foram construídos e testados os dispositivos estudados anteriormente, com o intuito de validar a aplicação desenvolvida. A ferramenta de simulação demonstrou ser capaz de simular de forma adequada os diversos aspectos do comportamento das LPG’s que foram estudados. A comparação dos resultados obtidos permitiu concluir que o modelo mais correcto para o estudo das LPG’s é o modelo de três camadas, o que está de acordo com o esperado.

Previsão do índice bovespa por meio de redes neurais artificiais: uma análise comparada aos métodos tradicionais de séries de tempo

Forecast is the basis for making strategic, tactical and operational business decisions. In financial economics, several techniques have been used to predict the behavior of assets over the past decades.Thus, there are several methods to assist in the task of time series forecasting, however, conventional modeling techniques such as statistical models and those based on theoretical mathematical models have produced unsatisfactory predictions, increasing the number of studies in more advanced methods of prediction. Among these, the Artificial Neural Networks (ANN) are a relatively new and promising method for predicting business that shows a technique that has caused much interest in the financial environment and has been used successfully in a wide variety of financial modeling systems applications, in many cases proving its superiority over the statistical models ARIMA-GARCH. In this context, this study aimed to examine whether the ANNs are a more appropriate method for predicting the behavior of Indices in Capital Markets than the traditional methods of time series analysis. For this purpose we developed an quantitative study, from financial economic indices, and developed two models of RNA-type feedfoward supervised learning, whose structures consisted of 20 data in the input layer, 90 neurons in one hidden layer and one given as the output layer (Ibovespa). These models used backpropagation, an input activation function based on the tangent sigmoid and a linear output function. Since the aim of analyzing the adherence of the Method of Artificial Neural Networks to carry out predictions of the Ibovespa, we chose to perform this analysis by comparing results between this and Time Series Predictive Model GARCH, developing a GARCH model (1.1).Once applied both methods (ANN and GARCH) we conducted the results' analysis by comparing the results of the forecast with the historical data and by studying the forecast errors by the MSE, RMSE, MAE, Standard Deviation, the Theil's U and forecasting encompassing tests. It was found that the models developed by means of ANNs had lower MSE, RMSE and MAE than the GARCH (1,1) model and Theil U test indicated that the three models have smaller errors than those of a naïve forecast. Although the ANN based on returns have lower precision indicator values than those of ANN based on prices, the forecast encompassing test rejected the hypothesis that this model is better than that, indicating that the ANN models have a similar level of accuracy . It was concluded that for the data series studied the ANN models show a more appropriate Ibovespa forecasting than the traditional models of time series, represented by the GARCH model

Análise fractal da vasculatura arteriolar e venular da retina em pessoas sem doença ocular

Although it has been suggested that retinal vasculature is a diffusion-limited aggregation (DLA) fractal, no study has been dedicated to standardizing its fractal analysis . The aims of this project was to standardize a method to estimate the fractal dimensions of retinal vasculature and to characterize their normal values; to determine if this estimation is dependent on skeletization and on segmentation and calculation methods; to assess the suitability of the DLA model and to determine the usefulness of log-log graphs in characterizing vasculature fractality . To achieve these aims, the information, mass-radius and box counting dimensions of 20 eyes vasculatures were compared when the vessels were manually or computationally segmented; the fractal dimensions of the vasculatures of 60 eyes of healthy volunteers were compared with those of 40 DLA models and the log-log graphs obtained were compared with those of known fractals and those of non-fractals. The main results were: the fractal dimensions of vascular trees were dependent on segmentation methods and dimension calculation methods, but there was no difference between manual segmentation and scale-space, multithreshold and wavelet computational methods; the means of the information and box dimensions for arteriolar trees were 1.29. against 1.34 and 1.35 for the venular trees; the dimension for the DLA models were higher than that for vessels; the log-log graphs were straight, but with varying local slopes, both for vascular trees and for fractals and non-fractals. This results leads to the following conclusions: the estimation of the fractal dimensions for retinal vasculature is dependent on its skeletization and on the segmentation and calculation methods; log-log graphs are not suitable as a fractality test; the means of the information and box counting dimensions for the normal eyes were 1.47 and 1.43, respectively, and the DLA model with optic disc seeding is not sufficient for retinal vascularization modeling

Análise e desenvolvimento de controladores preditivos multivariáveis baseados em multi-modelos bilineares

This work addresses issues related to analysis and development of multivariable predictive controllers based on bilinear multi-models. Linear Generalized Predictive Control (GPC) monovariable and multivariable is shown, and highlighted its properties, key features and applications in industry. Bilinear GPC, the basis for the development of this thesis, is presented by the time-step quasilinearization approach. Some results are presented using this controller in order to show its best performance when compared to linear GPC, since the bilinear models represent better the dynamics of certain processes. Time-step quasilinearization, due to the fact that it is an approximation, causes a prediction error, which limits the performance of this controller when prediction horizon increases. Due to its prediction error, Bilinear GPC with iterative compensation is shown in order to minimize this error, seeking a better performance than the classic Bilinear GPC. Results of iterative compensation algorithm are shown. The use of multi-model is discussed in this thesis, in order to correct the deficiency of controllers based on single model, when they are applied in cases with large operation ranges. Methods of measuring the distance between models, also called metrics, are the main contribution of this thesis. Several application results in simulated distillation columns, which are close enough to actual behaviour of them, are made, and the results have shown satisfactory

Um sistema para estimação da vazão de gás de poços produzindo por Plunger Lift para vaso separador de teste em plataformas de petróleo

This work intends to analyze the behavior of the gas flow of plunger lift wells producing to well testing separators in offshore production platforms to aim a technical procedure to estimate the gas flow during the slug production period. The motivation for this work appeared from the expectation of some wells equipped with plunger lift method by PETROBRAS in Ubarana sea field located at Rio Grande do Norte State coast where the produced fluids measurement is made in well testing separators at the platform. The oil artificial lift method called plunger lift is used when the available energy of the reservoir is not high enough to overcome all the necessary load losses to lift the oil from the bottom of the well to the surface continuously. This method consists, basically, in one free piston acting as a mechanical interface between the formation gas and the produced liquids, greatly increasing the well s lifting efficiency. A pneumatic control valve is mounted at the flow line to control the cycles. When this valve opens, the plunger starts to move from the bottom to the surface of the well lifting all the oil and gas that are above it until to reach the well test separator where the fluids are measured. The well test separator is used to measure all the volumes produced by the well during a certain period of time called production test. In most cases, the separators are designed to measure stabilized flow, in other words, reasonably constant flow by the use of level and pressure electronic controllers (PLC) and by assumption of a steady pressure inside the separator. With plunger lift wells the liquid and gas flow at the surface are cyclical and unstable what causes the appearance of slugs inside the separator, mainly in the gas phase, because introduce significant errors in the measurement system (e.g.: overrange error). The flow gas analysis proposed in this work is based on two mathematical models used together: i) a plunger lift well model proposed by Baruzzi [1] with later modifications made by Bolonhini [2] to built a plunger lift simulator; ii) a two-phase separator model (gas + liquid) based from a three-phase separator model (gas + oil + water) proposed by Nunes [3]. Based on the models above and with field data collected from the well test separator of PUB-02 platform (Ubarana sea field) it was possible to demonstrate that the output gas flow of the separator can be estimate, with a reasonable precision, from the control signal of the Pressure Control Valve (PCV). Several models of the System Identification Toolbox from MATLAB® were analyzed to evaluate which one better fit to the data collected from the field. For validation of the models, it was used the AIC criterion, as well as a variant of the cross validation criterion. The ARX model performance was the best one to fit to the data and, this way, we decided to evaluate a recursive algorithm (RARX) also with real time data. The results were quite promising that indicating the viability to estimate the output gas flow rate from a plunger lift well producing to a well test separator, with the built-in information of the control signal to the PCV

Controle preditivo bilinear aplicado a um motor de indução

The present work is based on the applied bilinear predictive control applied to an induction motor. As in particular case of the technique based on predictive control in nonlinem systems, these have desperted great interest, a time that present the advantage of being simpler than the non linear in general and most representative one than the linear one. One of the methods, adopted here, uses the linear model "quasi linear for step of time" based in Generalized Predictive Control. The modeling of the induction motor is made by the Vectorial control with orientation given for the indirect rotor. The system is formed by an induction motor of 3 cv with rotor in squirregate, set in motion for a group of benches of tests developed for this work, presented resulted for a variation of +5% in the value of set-point and for a variation of +10% and -10% in the value of the applied nominal load to the motor. The results prove a good efficiency of the predictive bilinear controllers, then compared with the linear cases

Estimação da demanda de instalações elétricas de edificações com múltiplas unidades de consumo

This work aims to predict the total maximum demand of a transformer that will be used in power systems to attend a Multiple Unit Consumption (MUC) in design. In 1987, COSERN noted that calculation of maximum total demand for a building should be different from that which defines the scaling of the input protection extension in order to not overestimate the power of the transformer. Since then there have been many changes, both in consumption habits of the population, as in electrical appliances, so that this work will endeavor to improve the estimation of peak demand. For the survey, data were collected for identification and electrical projects in different MUCs located in Natal. In some of them, measurements were made of demand for 7 consecutive days and adjusted for an integration interval of 30 minutes. The estimation of the maximum demand was made through mathematical models that calculate the desired response from a set of information previously known of MUCs. The models tested were simple linear regressions, multiple linear regressions and artificial neural networks. The various calculated results over the study were compared, and ultimately, the best answer found was put into comparison with the previously proposed model

Avaliacao de dados de secagem de suspensoes de polpas de frutas em leito de jorro com alimentacao intermitente

The objective of this work was the development and improvement of the mathematical models based on mass and heat balances, representing the drying transient process fruit pulp in spouted bed dryer with intermittent feeding. Mass and energy balance for drying, represented by a system of differential equations, were developed in Fortran language and adapted to the condition of intermittent feeding and mass accumulation. Were used the DASSL routine (Differential Algebraic System Solver) for solving the differential equation system and used a heuristic optimization algorithm in parameter estimation, the Particle Swarm algorithm. From the experimental data food drying, the differential models were used to determine the quantity of water and the drying air temperature at the exit of a spouted bed and accumulated mass of powder in the dryer. The models were validated using the experimental data of drying whose operating conditions, air temperature, flow rate and time intermittency, varied within the limits studied. In reviewing the results predicted, it was found that these models represent the experimental data of the kinetics of production and accumulation of powder and humidity and air temperature at the outlet of the dryer

Especificação do tipo intervalar parametrizado em CASL

The interval datatype applications in several areas is important to construct a interval type reusable, i.e., a interval constructor can be applied to any datatype and get intervals this datatype. Since the interval is, of certain form, a set of elements limited for two bounds, left and right, with a order notions, then it s reasonable that interval constructor enclose datatypes with partial order. On the order hand, what we want is work with interval of any datatype like this we work with this datatype then. it s important to guarantee the properties of the datatype when maps to interval of this datatype. Thus, the interval constructor get a theory to parametrized interval type, i.e., a interval with generics parameters (for example rational, real, complex). Sometimes, the interval application in some algebras doesn t guarantee the mainutenance of their properties, for example, when we use interval of real, that satisfies the field properties, it doesn t guarantee the distributivity propertie. A form to surpass this problem Santiago introduced the local equality theory that weakened the notion of strong equality, and thus, allowing some properties are local keeped, what can be discard before. The interval arithmetic generalization aim to apply the interval constructor on ordered algebras weakened for local equality with the purpose of the keep their properties. How the intervals are important in applications with continuous data, it s interesting specify that theory using a specification language that supply a system development using intervals of form disciplined, trustworth and safe. Currently, the algebraic specification language, based in math models, have been use to that intention often. We choose CASL (Common Algebraic Specification Language) among others languages because CASL has several characteristics excellent to parametrized interval type, such as, provide parcialiy and parametrization

Métodos estatísticos recursivos aplicados ao problema de estimação de vazão de gás em plantas de Plunger-lift

This work has as main objective to find mathematical models based on linear parametric estimation techniques applied to the problem of calculating the grow of gas in oil wells. In particular we focus on achieving grow models applied to the case of wells that produce by plunger-lift technique on oil rigs, in which case, there are high peaks in the grow values that hinder their direct measurement by instruments. For this, we have developed estimators based on recursive least squares and make an analysis of statistical measures such as autocorrelation, cross-correlation, variogram and the cumulative periodogram, which are calculated recursively as data are obtained in real time from the plant in operation; the values obtained for these measures tell us how accurate the used model is and how it can be changed to better fit the measured values. The models have been tested in a pilot plant which emulates the process gas production in oil wells

Modelos não lineares, alometria e dinâmica de macromoléculas para analisar o crescimento do Acará Bandeira (Pterophyllum scalare)

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Sensor em fibra óptica aplicado à caracterização de atuadores piezoelétricos flextensionais

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Modos deslizantes discretos em sistemas incertos com atraso na computação do sinal de controle

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)