This thesis focuses on two fundamental topics, specifically in medical statistics: the modelling of correlated survival datasets and the variable selection of the significant covariates and random effects. In particular, two types of survival data are considered: the classical survival datasets, where subjects are likely to experience only one type of event and the competing risks datasets, where subjects are likely to experience one of several types of event. In Chapter 2, among other topics, we highlight the importance of adding frailty terms on the proposed models in order to account for the association between the survival time and characteristics of subjects/groups. The main novelty of this thesis is to simultaneously select fixed effects and frailty terms through the proposed statistical models for each survival dataset. Chapter 3 covers the analysis of the classical survival dataset through the proposed Cox Proportional Hazard (PH) model. Utilizing a Cox PH frailty model, may increase the dimension of variable components and estimation of the unknown coefficients becomes very challenging. The method proposed for the analysis of classical survival datasets involves simultaneous variable selection on both fixed effects and frailty terms through penalty functions. The benefit of penalty functions is that they identify the non-significant parameters and set them to have a zero effect in the model. Hence, the idea is to 'doubly-penalize' the partial likelihood of the Cox PH frailty model; one penalty for each term. Estimation and selection implemented through Newton-Raphson algorithms, whereas closed iterative forms for the estimation and selection of fixed effects and prediction of frailty terms were obtained. For the selection of frailty terms, penalties imposed on their variances since frailties are random effects. Based on the same idea, we further extend the simultaneous variable selection in the competing risks datasets in Chapter 4, using extended cause-specific frailty models. Two different scenarios are considered for frailty terms; in the first case we consider that frailty terms vary among different types of events (similar to the fixed effects) whereas in the second case we consider shared frailties over all the types of events. Moreover, our 'individual penalization' approach allows for one covariate to be significant for some types of events, in contrast to the frequently used 'group-penalization' where a covariate is entirely removed when it is not significant over all the events. For both proposed methods, simulation studies were conduced and showed that the proposed procedure followed for each analysis works well in simultaneously selecting and estimating significant fixed effects and frailty terms. The proposed methods are also applied to real datasets analysis; Kidney catheter infections, Diabetes Type 2 and Breast Cancer datasets. Association of the survival times and unmeasured characteristics of the subjects was studied as well as a variable selection for fixed effects and frailties implemented successfully.

The main contribution of this thesis is to derive the properties and present a closed from solution of the exotic options under some specific types of Levy processes, such as American put options, American call options, British put options, British call options and American knock-out put options under either double exponential jump-diffusion processes or one-sided exponential jump-diffusion processes. Compared to the geometric Brownian motion, exponential jump-diffusion processes can better incorporate the asymmetric leptokurtic features and the volatility smile observed from the market. Pricing the option with early exercise feature is the optimal stopping problem to determine the optimal stopping time to maximize the expected options payoff. Due to the Markovian structure of the underlying process, the optimal stopping problem is related to the free-boundary problem consisting of an integral differential equation and suitable boundary conditions. By the local time-space formula for semi-martingales, the closed form solution for the options value can be derived from the free-boundary problem and we characterize the optimal stopping boundary as the unique solution to a nonlinear integral equation arising from the early exercise premium (EEP) representation. Chapter 2 and Chapter 3 discuss American put options and American call options respectively. When pricing options with early exercise feature under the double exponential jump-diffusion processes, a non-local integral term will be found in the infinitesimal generator of the underlying process. By the local time-space formula for semi-martingales, we show that the value function and the optimal stopping boundary are the unique solution pair to the system of two integral equations. The significant contributions of these two chapters are to prove the uniqueness of the value function and the optimal stopping boundary under less restrictive assumptions compared to previous literatures. In the degenerate case with only one-sided jumps, we find that the results are in line with the geometric Brownian motion models, which extends the analytical tractability of the Black-Scholes analysis to alternative models with jumps. In Chapter 4 and Chapter 5, we examine the British payoff mechanism under one-sided exponential jump-diffusion processes, which is the first analysis of British options for process with jumps. We show that the optimal stopping boundaries of British put options with only negative jumps or British call options with only positive jumps can also be characterized as the unique solution to a nonlinear integral equation arising from the early exercise premium representation. Chapter 6 provides the study of American knock-out put options under negative exponential jump-diffusion processes. The conditional memoryless property of the exponential distribution enables us to obtain an analytical form of the arbitrage-free price for American knock-out put options, which is usually more difficult for many other jump-diffusion models.

A nova is a cataclysmic variable star consisting of a white dwarf and a main sequence, subgiant or red giant companion. In a nova explosion, the white dwarf undergoes a thermonuclear runaway on its surface, as a result of build-up of accreted material from its companion. This leads to a large expulsion of matter from the WD surface, as well as a dramatic increase in the optical magnitude of the system. In this thesis, 5GHz C-band e-MERLIN observations of two novae (V959 Mon and V339 Del) and one symbiotic star (AG Pegasi) are presented. V959 Mon is a classical nova which was discovered in June 2012 by the Fermi Large Area Telescope as a gamma-ray source. High resolution e-MERLIN images of V959 Mon were made at six epochs following the June 2012 outburst, between 90 and 615 days after the Fermi discovery. The first four e-MERLIN observations revealed a morphology which was aspherical and expanding in the east-west direction. However, in the last two epochs, V959 Mon's ejecta morphology changed from being elongated east-west to being elongated north-south. Constant velocity models fit to measurements of V959 Mon's angular size at each epoch indicated that the expansion rate of the north-south component was significantly slower than that of the east-west component. The e-MERLIN observations of V339 Del and AG Pegasi were not as well resolved as the observations of V959 Mon, and only limited analysis of their ejecta structure was possible. In order to understand the complicated morphology seen in the e-MERLIN observations of V959 Mon, radio emission models of nova ejecta were constructed, and e-MERLIN observations of them were simulated. When constructing the models, two possible explanations for V959 Mon's morphology were explored. Firstly, the possibility was considered that the elongation observed in V959 Mon's ejecta, which changed over time from east- west to north-south, was spurious, resulting from incomplete sampling of the uv plane by e-MERLIN. To investigate this, an expanding spherical shell of ejecta was simulated, the extent to which its shape was distorted in simulated e-MERLIN images of it was investigated. Secondly, emission from an ejecta model featuring two components - one fast-moving component in the east-west direction, and another, more slow-moving component in the north-south direction - was simulated, in order to determine whether such an ejecta structure could result in the morphology seen in the e-MERLIN observations of V959 Mon. It was found that the spherical, Hubble flow model simulations were a reasonably good fit to V959 Mon's light curve, but could not explain the east-west elongation seen in V959 Mon's ejecta at the early epochs. Simulated observations of the two-component model were morphologically similar to V959 Mon, but produced light curves which were much fainter than V959 Mon's, indicating that a more in-depth analysis of two-component models of nova ejecta is necessary.

It is believed that structures and actuation systems should be tightly integrated together in the future to create fast moving, efficient, lightweight dynamic machines. Such actuated structures could be used for morphing aircraft wings, lightweight actuated space structures, or in robotics. This requires actuators to be distributed through the structure. A tensegrity structure is a very promising candidate for this future integration due to its potentially excellent stiffness and strength-to-weight ratio, and the inherent advantage of being a multi-element structure into which actuators can be embedded. Development of these machines will utilise expertise in several fields, involving kinematics, dynamics, actuation and multi-axis motion control. The research presented in this thesis concerns the study of multi-axis actuated tensegrity structures. A form-finding method has been developed to find stable geometries and determine stiffness properties of the type of tensegrity structure proposed. It has been shown that a tensegrity structure, with practical nodes of finite size, can be designed with actuated members to give shape-changing properties while potentially allowing a good stiffness to mass ratio. An antagonistic multi-axis control scheme has been developed for the tensegrity structure. The describing function technique has been used to analyse the dead band controller in the control scheme, giving a stability criterion. An experimental actuated tensegrity system has been designed and built incorporating pneumatic muscles controlled by switching valves. Mathematical models for the experimental actuated tensegrity system have been developed in detail, including the pneumatic actuation system and the structure geometry. The dynamic behaviour of the tensegrity system has been investigated via several simulation studies, using the developed models and the proposed control scheme. Experimental validation has been successfully conducted. The multi-axis control scheme can accurately control the tensegrity structure to achieve shape changes while maintaining a desired level of internal pre-load. The mathematical models can be used as a basis for further development.

Driven by the strict fuel consumption and CO2 legislations in Europe and many countries, various technologies have been developed to improve the fuel economy of conventional internal combustion engines. Gasoline engine downsizing has become a popular and effective approach to reduce fleet CO2 emissions of passenger cars. This is typically achieved in the form of boosted direct injection gasoline engines equipped with variable valve timing devices. Downsized gasoline engines reduce vehicle fuel consumption by making engine operate more at higher load to reduce pumping losses and also through reducing total engine friction losses. However, their compression ratio (CR) and efficiency are constrained by knocking combustion as well as the low speed pre-ignition phenomena. Miller cycle is typically achieved in an engine with reduced effective CR through Early Intake Valve Closure (EIVC) or Later Intake Valve Closure (LIVC). This technology has been adopted on modern gasoline engines to reduce in-cylinder charge temperature and enable a higher geometric CR to be used for better fuel economy. The present work investigated the effectiveness and underlying process of a Miller cycle based approach for improving fuel consumption of a boosted downsized gasoline engine. A single cylinder direct injection gasoline engine and the testing facilities were set up and used for extensive engine experiments. Both EIVC and LIVC approaches were tested and compared to the conventional Otto cycle operation with a standard cam profile. Synergy between Miller cycle valve timings and different valve overlap period was analysed. Two pistons with different CRs were used in the Miller cycle engine testing to enable its full potential to be evaluated. The experimental study was carried out in a large engine operation area from idle to up to 4000rpm and 25.6bar NIMEP to determine the optimal Miller cycle strategy for improved engine fuel economy in real applications. In addition, the increased exhaust back pressure and friction losses corresponding to real world boosting devices were calculated to evaluate Miller cycle benefits at high loads in a production engine. The results have shown that EIVC combined with high CR can offer up to 11% reduction of fuel consumption in a downsized gasoline engine with simple setup and control strategy. At the end, this thesis presents an Miller cycle based approach for maximising fuel conversion efficiency of a gasoline engine by combining three-stage cam profiles switching and two-stage variable compression ratio.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1981. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING. / Bibliography: leaves 109-110. / by Tim Emmanuel Bliamptis. / M.S.

O objetivo central desta tese é demonstrar a existência de uma estrutura regenerativa visível para cadeias de alcance variável não limitadas. Também apresentamos um algoritmo de identificação de seqüências de instantes de regeneração que converge quase certamente quando o tamanho da amostra diverge. / Our main aim is prove the existence of a regeneration scheme in unbounded variable length chains. We present an algorithm to identify sequences of regeneration times which converges almost surely as the sample length.

nÃo hà / The present study tries to expose a statistical analysis of the data collected from the Comptroller General of Discipline relating to disciplinary procedures applied to military police. In this sense, aims to qualitatively evaluate the statistical results obtained. The work developed correiÃÃo is critical to the performance of military police, considering, discouraging illicit acts by security agents and ensure a welcoming society and security against misconduct of security agents. In our study object the best option was a regression model censored tobit model that was the best fit to the study of variables collected. In the end it describes a qualitative analysis of the results for the variables found. / O presente estudo procura expor uma anÃlise estatÃstica dos dados coletados da Controladoria Geral de Disciplina, referente aos procedimentos disciplinares aplicados aos policiais militares. Neste sentido, visa avaliar qualitativamente os resultados estatÃsticos obtidos. O trabalho de correiÃÃo desenvolvido à de fundamental importÃncia para o desempenho dos policiais militares, haja vista, desencorajar a prÃtica de atos ilÃcitos por agentes de seguranÃa e garantir a sociedade um acolhimento e seguranÃa contra desvios de conduta dos agentes de seguranÃa. Na busca por alcanÃar o objetivo deste estudo, a melhor opÃÃo foi um modelo de regressÃo censurada, o modelo tobit, que foi o que melhor se adequou ao estudo das variÃveis colhidas. No final buscou-se descrever uma anÃlise qualitativa dos resultados referentes Ãs variÃveis encontradas.

A seleção de variáveis é uma importante fase para a construção de um modelo parcimonioso. Entretanto, as técnicas mais populares de seleção de variáveis, como, por exemplo, a seleção do melhor subconjunto de variáveis e o método stepwise, ignoram erros estocásticos inerentes à fase de seleção das variáveis. Neste trabalho, foram estudados procedimentos alternativos aos métodos mais populares para o modelo de riscos proporcionais de Cox e o modelo de Cox com fragilidade gama. Os métodos alternativos são baseados em verossimilhançaa penalizada e diferem dos métodos usuais de seleção de variáveis, pois têm como objetivo excluir do modelo variáveis não significantes estimando seus coeficientes como zero. O estimador resultante possui propriedades desejáveis com escolhas apropriadas de funções de penalidade e do parâmetro de suavização. A avaliação desses métodos foi realizada por meio de simulação e uma aplicação a um conjunto de dados reais foi considerada. / Variable selection is an important step when setting a parsimonious model. However, the most popular variable selection techniques, such as the best subset variable selection and the stepwise method, do not take into account inherent stochastic errors in the variable selection step. This work presents a study of alternative procedures to more popular methods for the Cox proportional hazards model and the frailty model. The alternative methods are based on penalized likelihood and differ from the usual variable selection methods, since their objective is to exclude from the model non significant variables, estimating their coefficient as zero. The resulting estimator has nice properties with appropriate choices of penalty functions and the tuning parameter. The assessment of these methods was studied through simulations, and an application to a real data set was considered.

Neste trabalho foi realizado um estudo de aspectos relacionados ao uso de uma Rede de Sensores sem Fio, como uma tecnologia de apoio para o monitoramento ambiental em vinhedos. Com o intuito de estender o tempo de duração das baterias dos nos sensores dessa rede, procurou-se evitar coletas desnecessárias de dados. O programa aplicativo dos nos foi redefinido para possibilitar a escolha de um intervalo de amostragem próximo das necessidades, sendo essa escolha feita pelos próprios nos, a partir de condições predefinidas no aplicativo. A área de aplicação escolhida foi a viticultura, devido a grande influencia das condições ambientais no desenvolvimento fisiológico das videiras, influenciando no rendimento e na qualidade das uvas produzidas, justificando a adoção de um sistema de monitoramento. Uma das características dessas redes que motiva a sua aplicação na área agrícola e a possibilidade de coleta de dados com a resolução espaço-temporal adequada, permitindo aplicar eficientemente os conceitos da Agricultura de Precisão. Realizaram-se experimentos em um parreiral cultivado sob cobertura plástica, localizado em Bento Gonçalves, RS. Os aspectos da rede estudados compreenderam: o alcance de conectividade e qualidade do enlace entre os nos e a estação base; o desempenho da rede com os nos posicionados em diferentes alturas; a exatidão dos dados coletados, em comparação com um equipamento de maior precisão; e a avaliação da variabilidade microclimática dentro do parreiral. A validação da proposta envolveu a implementação de um protótipo funcional, testado em laboratório. O modelo proposto permitiu estimar o tempo de vida dos nos, considerando-se parâmetros como o tipo de bateria, o intervalo de amostragem e o tamanho do pacote de dados enviados pelos nos. Os resultados obtidos evidenciaram a existência de variabilidade microclimática tanto em diferentes alturas da videira, como espacialmente, dentro do parreiral. A interferência dos vegetais devido a água que contém foi avaliada a partir de dois parâmetros: o indicador de intensidade de sinal recebido e de qualidade de enlace, que indicaram maior interferência na altura do dossel das videiras. A correlação entre esses parâmetros levou a constatação que o melhor desempenho e o maior alcance de conectividade ocorrem com os nos posicionados na altura do cacho das videiras se comparados com as medições realizadas no dossel. A amostragem variável permitiu uma economia no numero de pacotes enviados, com menor consumo de energia, o que foi comprovado com o modelo proposto para tal finalidade. / A study of issues related to using a Wireless Sensor Network was carried out, as an enabling technology for environmental monitoring in vineyards. In order to extend the duration of sensor nodes batteries of the network, collecting unnecessary data was avoided. The nodes application program was redefined to allow the choice of a sampling interval close to the needs, and the choice made by the nodes themselves, based on predefined conditions in their application. The application area chosen was the viticulture because of the large influence of environmental conditions on the physiological development of the vines, influencing the yield and quality of produced grapes, justifying the adoption of a monitoring system. One of the characteristics of such networks that motivates its application in agriculture is the possibility of collecting data with the appropriate spatial and temporal resolution, allowing to efficiently apply the concepts of Precision Agriculture. Experiments were conducted in a vineyard cultivated under plastic overhead cover, situated in Bento Gonçalves, RS. The network aspects studied involved: connectivity range and link quality between nodes and gateway; network performance with nodes positioned at different heights; collected data accuracy, compared with a higher precision equipment; and microclimatic variability evaluation within the vineyard. The proposal validation involved the implementation of a functional prototype, tested in the laboratory. The model allowed estimating the nodes lifetime, considering parameters such as battery type, sampling interval and data packet size sent by sensor nodes. Results showed the existence of microclimatic variability both at different heights of the vine, and spatially within the vineyard. Vegetables interference due to the water contained in them was evaluated from two parameters: Received Signal Strength Indicator and Link Quality Indicator, which indicated greater interference in the height of the vines canopy. The correlation between these parameters led to the conclusion that the best performance and increased connectivity range occur with nodes positioned at the height of the bunch of grapes as compared with measurements made in the canopy. The variable sampling allowed savings in the number of sent packets, with lower power consumption, which agrees with the proposed model for this purpose.