Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-03-02T13:24:53Z No. of bitstreams: 1 guilhermeguilherminoneto.pdf: 1186805 bytes, checksum: c3c46b18aeea88b1e1d4b4fa3d1a36e5 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-03-06T19:39:48Z (GMT) No. of bitstreams: 1 guilhermeguilherminoneto.pdf: 1186805 bytes, checksum: c3c46b18aeea88b1e1d4b4fa3d1a36e5 (MD5) / Made available in DSpace on 2017-03-06T19:39:48Z (GMT). No. of bitstreams: 1 guilhermeguilherminoneto.pdf: 1186805 bytes, checksum: c3c46b18aeea88b1e1d4b4fa3d1a36e5 (MD5) Previous issue date: 2014-08-20 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Previsões de demanda em curto prazo são fundamentais para o planejamento e o controle da produção em sistemas de energia elétrica. Como não é viável manter estoques de segurança para compensar demandas inesperadas, a programação da geração é baseada em previsões feitas com antecedência de algumas horas. Ao longo dos anos, muitos métodos foram testados para a resolução do problema. Dentre os mais populares estão os univariados, em que a demanda é escrita como uma função linear de seu comportamento histórico e prevista por técnicas estatísticas. Também é frequente o uso de métodos multivariados, que levam em conta o efeito não- linear de variáveis climáticas, como a temperatura do ar, sobre o comportamento do consumidor. Para este caso, a literatura recente sugere o uso de previsores de inteligência computacional, como as redes neurais artificiais. Embora alguns autores afirmem que deve-se considerar métodos multivariados, outros defendem que, para previsões de curto prazo (horizonte de poucas horas), a inclusão de variáveis climáticas traz poucos benefícios, posto que seus efeitos levam mais tempo para serem percebidos. Neste trabalho, experimentamos diversos métodos univariados e multivariados a fim de comparar seu desempenho sobre uma base de dados da cidade do Rio de Janeiro. Para estes dados, mostramos que é possível obter, por meio de um simples previsor linear univariado (um modelo de curva de carga cuja componente-padrão é prevista pelo amortecimento de Holt-Winters-Taylor), resultados próximos aos de técnicas mais complexas, porém, com as vantagens de maior robustez, parcimônia e economia de recursos computacionais. / Short-term demand forecasts a are vital part of the production plan and control on electrical power systems. As it is not possible to keep large inventories to meet sudden demand increases, the generation scheduling is based on forecasts made for some hours ahead. Throughout the years, many methods have been proposed in order to solve the problem. Among the most popular are the univariate ones, on which the demand is written as a linear function of its historical behavior and forecast by statistical techniques. It is also common to use multivariate methods, which take into account also the nonlinear effects produced on the demand by weather-related variables, such as the air temperature. For this case, recent papers suggest the use of computational intelligence devices, such as artificial neural networks. Although some authors claim that multivariate methods must be considered, some others state that, on a short-run (lead-times up to a few hours), adding weather-related variables brings little benefits, because its effects might take a longer time to affect the demand. On this work, we experiment a large amount of univariate and multivariate methods aiming to compare its performance over a dataset from the city of Rio de Janeiro. For these data, we show that is possible to obtain, via a simple linear univariate method (a load curve model where the standard load is forecast by the Holt-Winters-Taylor smoothing), results that are close enough to those achieved by more complex techniques, but bringing the advantages of more robustness, parsimony and computational economy.

Submitted by Geandra Rodrigues (geandrar@gmail.com) on 2018-10-24T12:50:06Z No. of bitstreams: 1 taisdemedeirospedreira.pdf: 1862768 bytes, checksum: 0c6ee31fd9be772b5b609051a207f61f (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2018-11-23T12:17:19Z (GMT) No. of bitstreams: 1 taisdemedeirospedreira.pdf: 1862768 bytes, checksum: 0c6ee31fd9be772b5b609051a207f61f (MD5) / Made available in DSpace on 2018-11-23T12:17:19Z (GMT). No. of bitstreams: 1 taisdemedeirospedreira.pdf: 1862768 bytes, checksum: 0c6ee31fd9be772b5b609051a207f61f (MD5) Previous issue date: 2018-09-05 / As previsões a curto prazo da carga elétrica (de algumas horas até alguns dias à frente) são essenciais para o planejamento, controle e operação dos sistemas de energia, tanto por por razões técnicas quanto financeiras. Como não é possível estocar grandes quantidades, torna-se indispensável um maneira eficaz de programar a produção da energia para que ela atenda a demanda. Por conta disso, uma grande literatura desenvolveu-se sobre o assunto. Devido à complexidade das séries de carga e à dependência não-linear destas carga em relação a diversas variáveis exógenas, os sistemas de previsão mais frequentemente propostos em trabalhos recentes são aqueles baseados em algoritmos complexos de inteligência computacional. No entanto, métodos lineares simples ainda são muito comumente usados, por si sós ou em combinação com técnicas não-lineares. Um desses métodos é o de Holt-Winters-Taylor, que é uma adaptação do conhecido método de amortecimento exponencial de Holt-Winters para que múltiplas sazonalidades possam ser modelados concomitantemente. Este trabalho implementa três variantes deste método HWT e analisa seus desempenhos em duas séries de dados reais de carga. Verificou-se que uma combinação linear dessas variantes nitidamente supera o método HWT original e fornece previsões precisas, com um baixo custo computacional. / Short-term load forecasts (forecasts for horizons ranging from a few hours to a few days ahead) are essential for the planning, controling and operation of energy systems, both for technical and financial reasons. Since it is not feasible to store energy in large quantities, an efficient way to forecast energy demand becomes indispensable. Because of this, a large literature has developed on the subject. Due to the complexity of load series and the nonlinear relationship of the load with exogenous variables, the most frequently proposed forecasting systems in recent papers are those based on complex algorithms of computational intelligence. However, simple linear methods are still very frequently used, either alone or in combination with non-linear techniques. One of these methods is Holt-Winters-Taylor (HWT), which is an adaptation of the well-known Holt-Winters exponential smoothing method, modified so that multiple seasonalities can be modeled at the same time. In this paper, we implement three variants of this HWT method and analyze their performances over two sets of actual load data. We found that a linear combination of these variants clearly outperforms the original HWT method, and provides accurate forecasts at a low computational cost.

En analyse multivariée de données de grande dimension, les lois de Wishart définies dans le contexte des modèles graphiques revêtent une grande importance car elles procurent parcimonie et modularité. Dans le contexte des modèles graphiques Gaussiens régis par un graphe G, les lois de Wishart peuvent être définies sur deux restrictions alternatives du cône des matrices symétriques définies positives : le cône PG des matrices symétriques définies positives x satisfaisant xij=0, pour tous sommets i et j non adjacents, et son cône dual QG. Dans cette thèse, nous proposons une construction harmonieuse de familles exponentielles de lois de Wishart sur les cônes PG et QG. Elle se focalise sur les modèles graphiques d'interactions des plus proches voisins qui présentent l'avantage d'être relativement simples tout en incluant des exemples de tous les cas particuliers intéressants: le cas univarié, un cas d'un cône symétrique, un cas d'un cône homogène non symétrique, et une infinité de cas de cônes non-homogènes. Notre méthode, simple, se fonde sur l'analyse sur les cônes convexes. Les lois de Wishart sur QAn sont définies à travers la fonction gamma sur QAn et les lois de Wishart sur PAn sont définies comme la famille de Diaconis- Ylvisaker conjuguée. Ensuite, les méthodes développées sont utilisées pour résoudre la conjecture de Letac- Massam sur l'ensemble des paramètres de la loi de Wishart sur QAn. Cette thèse étudie aussi les sousmodèles, paramétrés par un segment dans M, d'une famille exponentielle paramétrée par le domaine des moyennes M. / In the framework of Gaussian graphical models governed by a graph G, Wishart distributions can be defined on two alternative restrictions of the cone of symmetric positive definite matrices: the cone PG of symmetric positive definite matrices x satisfying xij=0 for all non-adjacent vertices i and j and its dual cone QG. In this thesis, we provide a harmonious construction of Wishart exponential families in graphical models. Our simple method is based on analysis on convex cones. The focus is on nearest neighbours interactions graphical models, governed by a graph An, which have the advantage of being relatively simple while including all particular cases of interest such as the univariate case, a symmetric cone case, a nonsymmetric homogeneous cone case and an infinite number of non-homogeneous cone cases. The Wishart distributions on QAn are constructed as the exponential family generated from the gamma function on QAn. The Wishart distributions on PAn are then constructed as the Diaconis- Ylvisaker conjugate family for the exponential family of Wishart distributions on QAn. The developed methods are then used to solve the Letac-Massam Conjecture on the set of parameters of type I Wishart distributions on QAn. Finally, we introduce and study exponential families of distributions parametrized by a segment of means with an emphasis on their Fisher information. The focus in on distributions with matrix parameters. The particular cases of Gaussian and Wishart exponential families are further examined.

Cette thèse montre la possibilité d’une application rigoureuse de la méthode d’intégrationnumérique double-exponentielle introduite par Takahasi et Morien 1974, et sa pertinence pour lescalculs à grande précision en théorie des nombres. Elle contient en particulier une étude détailléede cette méthode, des critères simples sur son champ d’application, et des estimations rigoureusesdes termes d’erreur.Des paramètres explicités et précis permettent de l’employer aisément pour le calcul garantide fonctions définies par des intégrales.Cette méthode est également appliquée en détail au calcul de transformées de Mellin inversesde facteurs gamma intervenant dans les calculs numériques de fonctions L. Par une étude unifiée,ce travail démontre la complexité d’un algorithme de M. Rubinstein et permet de proposer desalgorithmes de calcul de valeurs de fonctions L quelconques dont le résultat est garanti et dont lacomplexité est meilleure en la précision. / This thesis contains a detailed study of the so-called double exponential integration formulasintroduced by Takahasi and Moriin 1974,and provides explicit bounds forarigorous applicationof the method in number theory.Accurate parameters are given, which makes it possible to use it as a blackbox for the rigorouscomputation of functions defined by integrals.It also deals with numerical computations of L functions. The complexity of analgorithm dueto M. Rubinstein is proven. In the context of double-exponential transformation, a new algorithmis provided whose complexity is low in terms of precision.

No description available.

Manufacturing systems have developed both physically and technologically, allowing production of innovative new products in a shorter lead time, to meet the 21st century market demand. Flexible flow lines for instance use flexible entities to generate multiple product variants using the same routing. However, the variability within the flow line is asynchronous and stochastic, causing disruptions to the throughput rate. Current autonomous variability control approaches decentralise the autonomous decision allowing quick response in a dynamic environment. However, they have limitations, e.g., uncertainty that the decision is globally optimal and applicability to limited decisions. This research presents a novel formula-based autonomous control method centered on an empirical study of the effect of stochastic variability on the performance of flexible human-dependent serial flow lines. At the process level, normal distribution was used and generic nonlinear terms were then derived to represent the asynchronous variability at the flow line level. These terms were shortlisted based on their impact on the throughput rate and used to develop the formula using data mining techniques. The developed standalone formulas for the throughput rate of synchronous and asynchronous human-dependent flow lines gave steady and accurate results, higher than closest rivals, across a wide range of test data sets. Validation with continuous data from a real-world case study gave a mean absolute percentage error of 5%. The formula-based autonomous control method quantifies the impact of changes in decision variables, e.g., routing, arrival rate, etc., on the global delivery performance target, i.e., throughput, and recommends the optimal decisions independent of the performance measures of the current state. This approach gives robust decisions using pre-identified relationships and targets a wider range of decision variables. The performance of the developed autonomous control method was successfully validated for process, routing and product decisions using a standard 3x3 flexible flow line model and the real-world case study. The method was able to consistently reach the optimal decisions that improve local and global performance targets, i.e., throughput, queues and utilisation efficiency, for static and dynamic situations. For the case of parallel processing which the formula cannot handle, a hybrid autonomous control method, integrating the formula-based and an existing autonomous control method, i.e., QLE, was developed and validated.

>Magister Scientiae - MSc / HIV/AIDS is nowadays considered as the greatest public health disaster of modern time. Its progression has challenged the global population for decades. Through mathematical modelling, researchers have studied different interventions on the HIV pandemic, such as treatment, education, condom use, etc. Our research focuses on different compartmental models with emphasis on the effect of public health education. From the point of view of statistics, it is well known how the public health educational programs contribute towards the reduction of the spread of HIV/AIDS epidemic. Many models have been studied towards understanding the dynamics of the HIV/AIDS epidemic. The impact of ARV treatment have been observed and analysed by many researchers. Our research studies and investigates a compartmental model of HIV with treatment and education campaign. We study the existence of equilibrium points and their stability. Original contributions of this dissertation are the modifications on the model of Cai et al. [1], which enables us to use optimal control theory to identify optimal roll-out of strategies to control the HIV/AIDS. Furthermore, we introduce randomness into the model and we study the almost sure exponential stability of the disease free equilibrium. The randomness is regarded as environmental perturbations in the system. Another contribution is the global stability analysis on the model of Nyabadza et al. in [3]. The stability thresholds are compared for the HIV/AIDS in the absence of any intervention to assess the possible community benefit of public health educational campaigns. We illustrate the results by way simulation The following papers form the basis of much of the content of this dissertation, [1 ] L. Cai, Xuezhi Li, Mini Ghosh, Boazhu Guo. Stability analysis of an HIV/AIDS epidemic model with treatment, 229 (2009) 313-323. [2 ] C.P. Bhunu, S. Mushayabasa, H. Kojouharov, J.M. Tchuenche. Mathematical Analysis of an HIV/AIDS Model: Impact of Educational Programs and Abstinence in Sub-Saharan Africa. J Math Model Algor 10 (2011),31-55. [3 ] F. Nyabadza, C. Chiyaka, Z. Mukandavire, S.D. Hove-Musekwa. Analysis of an HIV/AIDS model with public-health information campaigns and individual with-drawal. Journal of Biological Systems, 18, 2 (2010) 357-375. Through this dissertation the author has contributed to two manuscripts [4] and [5], which are currently under review towards publication in journals, [4 ] G. Abiodun, S. Maku Vyambwera, N. Marcus, K. Okosun, P. Witbooi. Control and sensitivity of an HIV model with public health education (under submission). [5 ] P.Witbooi, M. Nsuami, S. Maku Vyambwera. Stability of a stochastic model of HIV population dynamics (under submission).

This Thesis deals with the power of a statistical test and the associated problem of determining the appropriate sample size. It should be large enough to meet the requirements of the probabilities of errors of both the first and second kind. The aim of this Thesis is to demonstrate theoretical methods that result in derivation of formulas for minimum sample size determination. For this Thesis, three important probability distributions have been chosen: Normal, Bernoulli, and Exponential.

In this Master Thesis there are summarized basic methods for modelling time series, such as linear regression with seasonal dummy variables, exponential smoothing and SARIMA processes. The thesis is aimed on modelling and forecasting seasonal time series using these methods. Goals of the Thesis are to introduce and compare these methods using a set of 2184 seasonal time series followed by evaluation their prediction abilities. The main benefit of this Master Thesis is understanding of different aspects of forecasting time series and empirical verification of advantages and disadvantages these methods in field of creating predictions.

No description available.