A presente pesquisa se refere à aplicação dos métodos estocásticos para estudar fenômenos críticos em modelos de sistemas classificados como desordenados que apresentam transição de fase do tipo-ordem desordem. Essa pesquisa é definida tanto no quadro teórico da Mecânica Estatística dos fenômenos críticos e transição de fase de equilíbrio e fora de equilibrio, com os recursos associados à análise de escala de tamanho finito quanto no quadro dos recursos aos processos estocásticos markovianos, descritos pela equação-mestra e associados a técnicas essencialmente numéricas como o método estocástico computacional de Monte Carlo. Na primeira etapa desta pesquisa, os modelos estudados são da classe denominada de votante majoritário. Eles são indexados pelo número z de vizinhos mais próximos com spin central, tem dois estados e são construidos em redes quadradas. A evolução dinâmica é dada pela regra da maioria junto com regradfe desempate. Eles não satisfazem a propriedade do principio do balanceamento detalhado, portanto, são classificados como descrevendo fenômenos fora do equilíbrio. Contudo, eles satisfazem a propriedade de simetria de inversão de sinal, o que os coloca teoricamente na classe de universalidade do modelo de Ising. Desta forma, a evolução dinâmica desses modelos é estudada com os recursos da equação mestra ou equação de evolução. No entanto, essa abordagem teórica é feita apenas na aproximação de campo médio, a qual fornece, na solução estacionária, os valores clássicos para os parametros relevantes. Em contrapartida, os valores numéricos exatos para os valores do ponto crítico e dos expoentes críticos, que são não clássicos, é dada por meio do recurso ao método de simulação computacional e à análise de escala de tamanho finito. Esses valores confirmam o resultado teórico quanto à classe de universalidade para cada modelo específico. Na sequência, estudam-se as propriedades dos modelos resultantes da combinação convexa do votante majoritário. Os resultados são semelhantes aos anteriores. Um resultado extra permitido por essas combinações convexa éa construção de uma relação contínua entre o valor crítico indutor da transição de fase e o número de vizinhos. Neste contexto foi apresentada uma solução para o problema do modelo mais simples desta classe de modelos. Com o modelo mais simples ilustram-se as condições universais de transição de fase, em particular o papel da dimensão do sistema. Na segunda etapa da pesquisa, cvonstrí-se, então, outra classe de modelos do votante que, por analogia com o modelo de Ising, tem como estado fundamental a fasse antiferromagnética: a classe dos modelos do votante minoritário. Essa classe de modelos possue as mesmas propriedades da classe de modelos do votante majoritário e por isso obtem-se os mesmos resultados. A analogia com o modelo de Ising é levada um pouco mais longe com a construção de um análogo aos modelo +-J: a construção da combinação convexa do votante majoritário com o minoritário. Para esse novo modelo constrói-se tanto o diagrama com as três fases, ferromagnética, paramagnética e antiferromagnética quanto as concentrações críticas que as distinguem. Não se obtém uma possível fase de vidro de spin. Uma vez que os modelos do votante são originalmente tidos como sistemas desordenados, comparam-se, para um mesmo modelo, resultados obtidos pela aplicação de dois diferentes métodos de tratar os modelos de sistemas desordenados: o método temperado-\"quenched\" - e o método recozido - \"annealed\". Na terceira etapa desta pesquisa e na mesma linha dos modelos estocásticos irreversíveis tratados anteriormente, estuda-se ainda outro modelo, classificado como jogo espacial nos sítios de uma rede quadrada. Simulações mostram que além de dois estados absorventes ha\'também a presença de um estado ativo definido por uma densidade finita de cooperadores e não cooperadores e que esse modelo se encontra na classe de universalidade do modelo de percolação direcionada. Nesta mesma etapa, mas, agora, no contexto da Mecânica sStatística de Equilíbrio, aborda-se o modelo de Ising quântico unidimensional com campo transverso por meio de simulação de Monte Carlo.Com o uso do método estiocástico e por meio da curva do colapso calculam-se os valores do ponto crítico e dos expoentes críticos desse modelo. / This research refers to the applications of the stochastic methods to the study of the critical phenomena in models of systems classified as disordered that undergo phase transition of the order-disorder kind. This research is defined as in the theoretical framework of the Statistical mechanics of the equilibrium and non-equilibrium of the critical phenomena and phase transition with the resources associated to the analysis of finite-size scale, as in the frame of the resources of markovian stochastic process described by the master equation associated with essentially numerical techniques such as stochastic computational method of Monte Carlo.l In this first stage of this research, the studied models belong to the class of the majority voter. They are described by a lattice with spins in each site with two states. The dynamic of these models is described by the majority rule together with a rule for solving problems of indecision. These models do not obey the principle of microscopic reversibility therefore they are classified as describing phenomena of non-equilibrium. However, they satisfy the property of \"up-down\" symmetry which make theoretically belong to the universality class of the Ising model. The mean field approach to the master equation is done and the exact value is pursued by the use of the method of the computational simulation with theuse of the analysis of finite-size scale. The results obtained for the critical exponents support the hypothesis of universality class of these models. There are constructions of the convex combination of these models. A question is raised about the simplest model and a possible solution is presented. There is a search for another kind of majority voter, but with an antiferromagnetic ground state, which leads to the minority voter. It is also to be classified in the same universality class. A natural unfold of this research is making the convex combination of the minority and majority voter models by analogy with the Ising model +- J and ask for the phase diagram class.Some results are also obtained by comparing the quenched and annealed approach to a same majority voter model. Finally, there are two more applications of these methods for obtaining critical point and critical exponents. The first refers to a model with absorbing state which is classified in the universality class of direct percolation. The second refers to a quantum model with transverse field.

Neste trabalho propomos o uso de um método Bayesiano para estimar o parâmetro de memória de um processo estocástico com memória longa quando sua função de verossimilhança é intratável ou não está disponível. Esta abordagem fornece uma aproximação para a distribuição a posteriori sobre a memória e outros parâmetros e é baseada numa aplicação simples do método conhecido como computação Bayesiana aproximada (ABC). Alguns estimadores populares para o parâmetro de memória serão revisados e comparados com esta abordagem. O emprego de nossa proposta viabiliza a solução de problemas complexos sob o ponto de vista Bayesiano e, embora aproximativa, possui um desempenho muito satisfatório quando comparada com métodos clássicos. / In this work we propose the use of a Bayesian method for estimating the memory parameter of a stochastic process with long-memory when its likelihood function is intractable or unavailable. Such approach provides an approximation for the posterior distribution on the memory and other parameters and it is based on a simple application of the so-called approximate Bayesian computation (ABC). Some popular existing estimators for the memory parameter are reviewed and compared to this method. The use of our proposal allows for the solution of complex problems under a Bayesian point of view and this proposal, although approximative, has a satisfactory performance when compared to classical methods.

Este trabalho tem por objetivo discutir as metodologias de estimação da PD utilizadas na indústria financeira. Além disso, contextualizar a aplicação do trabalho ao IFRS9 e seu direcionamento para o tema de Risco de Crédito. Historicamente os grandes bancos múltiplos utilizam variadas metodologias econométricas para modelar a Probabilidade de Descumprimento (PD),um dos métodos mais tradicionais é a regressão logística, entretanto com a necessidade do cálculo da Perda Esperada de Crédito através do IFRS9, se torna necessário mudar o paradigma de estimação para uma abordagem forward-looking, isto está sendo interpretado por muitas instituições e consultorias como a inclusão de fatores e variáveis projetadas dentro do processo de estimação, ou seja, não serão utilizados apenas os dados históricos para prever o descumprimento ou inadimplência. Dentro deste contexto será proposto uma abordagem que une a estimação da Probabilidade de Descumprimento com a inclusão de um fator foward-looking. / This paper aims to discuss the methodologies used to estimate the Probability Of Default used in the financial industry. In addition, contextualize the application of the work to IFRS9 requirements and its targeting to the Credit Risk theme. Historically large multi-banks use a variety of econometric methodologies to model the Probability of Default, one of the more traditional methods is logistic regression. However, with the need to calculate the expected credit loss through IFRS9, it becomes necessary to change the estimation paradigm to a forwardlooking approach, this is being interpreted by many institutions and consultancies companies as the inclusion of factors and variables projected within the estimation process, that is, not only historical data are used to predict the default. Within this context will be proposed an approach that joins the estimation of Probability of Default with the inclusion of a forward-looking factor.

Nous présentons dans cette thèse une méthode d’évaluation de l’efficience basée sur la modélisation temporelle et stochastique et de la simulation de la propagation des fautes dans les systèmes industriels. Les TRS (Taux de Rendement Synthétique) est devenu au travers de la norme NF E60-182 l’un des indicateurs majeurs de l’efficience du pilotage des systèmes de production. Il intègre essentiellement 3 notions (Qualité, Productivité et Disponibilité). Si son expression pour un composant est assez simple, sa modélisation pour l’évaluation du comportement fonctionnel et dysfonctionnel l’est beaucoup moins (prise en compte des facteurs d’échelle, des désynchronisations et du facteur humain). Afin de permettre une prise en compte des contributions individuelles de chaque composante du TRS et de chaque partie constitutive des systèmes de production (technique et humaine), nous avons utilisé les automates d’états comme haut langage de description. Les attendus débouchent en amont sur un apport formel pour l’établissement d’une méthodologie d’analyse et de conception, et en aval sur une fourniture d’indicateurs décisionnels. Les résultats sont implantés autour d’un démonstrateur basé sur AltaRica Data-Flow langage à la fois formel et graphique et véritable outil de modélisation / simulation. / We present in this thesis a method to assess a system efficiency based on modelling of the temporal and stochastic spread of faults. The OEE (Overall Equipment Effectiveness) has become through the NF E60-182 standard one of the major indicators of the effectiveness in piloting production systems. It includes 3 main concepts (Quality, Performance and Availability). If its evaluation can be easy for a single system, the modeling of its components for the global efficiency assessment is much more difficult (taking into account redundancies, temporal scale factors…).in order to take in account the local contribution of each component of complex production system, a notion of OTE (Overall Throughput Effectiveness) is developed. The purpose of OTE is twofold: it measures factory-level performance and factory-level diagnostics such as bottleneck detection. The expected result gives a formal contribution to establishment a methodology for analysis, design, and decision-making. The results are discussed using a demonstrator based on AltaRica Data-Flow, language in both formal and graphic and real tool modeling / simulation.

Uma série de problemas de filas em sistemas de transportes com picos de tráfego, ou um número finito de elementos no sistema, são usualmente representados pelo modelo de fonte infinita, dadas as dificuldades de utilização do modelo exato de fonte finita. Este trabalho apresenta uma solução alternativa baseada no cálculo de um limitante superior para as probabilidades de equilíbrio do modelo exato de fonte finita, e compara as medidas de desempenho dos sistemas calculadas pelos dois modelos. Mostra-se que para índices de congestionamento menores que um, as diferenças entre estas medidas são tanto menores quanto menor for este índice. A partir destes resultados, compara-se as medidas de desempenho do modelo aproximado proposto com as do modelo de população infinita, para diferentes tamanhos de população e números de servidores. Conclui-se que os modelos conduzem a resultados numéricos muito próximos para uma ampla variação do índice de congestionamento, e que estes resultados são tão melhores quanto maior for o número de servidores no sistema e o número de elementos na população. São também apresentados três estudos de casos comparando os resultados induzidos pelos modelos exato, proposto e usual de fonte infinita, que ilustram a aplicabilidade prática dos resultados deste trabalho em sistemas de transportes. / A set of stochastic queueing problems in transportation systems with traffic peaks, or a finite number of elements in the system, are usually represented by the infinite source model, due to the difficulties of applying the exact finite source model. This study presents an alternative solution based on the upper bound values of the equilibrium probabilities of the exact finite source model, and compares the performance measurements of the two models. It is shown that for congestion factors below one, the smaller the value of the congestion factor, the smaller is the difference between the models. Based on this results the measures of performance of the proposed aproximate model are compared with the results of the usual infinite source model for different population sizes and number of servers. It is concluded that the models lead to very close numerical results for a wide range of congestion factors of the system and that these results are the better the larger is the number of servers and the number of elements in the system. Three case studies ilustrating the pratical applicability of the results of this study to transportation systems are also presented.

Depuis le début des années 90, les automates temporisés et les langages temporisés ont été largement utilisés pour modéliser et vérifier les systèmes temps réels. Ces langages ont été aussi été largement étudiés d'un point de vue théorique. Plus récemment Asarin et Degorre ont introduit les notions de volume et d'entropie des langages temporisés pour quantifier la taille de ces langages et l'information que ses éléments contiennent. Dans cette thèse nous construisons de nouveaux développements à cette théorie (que nous appelons volumétrie des langages temporisés) et l'appliquons a plusieurs problèmes apparaissant dans divers domaine de recherche tel que la théorie de l'information, la vérification, la combinatoire énumérative. Entre autre nous (i) développons une théorie de la dynamique symbolique temporisée~; (ii) caractérisons une dichotomie entre automate temporisé se comportant bien ou mal~; (iii) définissons pour un automate temporisé donné, un processus stochastique d'entropie maximale le moins biaisé possible~; (iv) développons une version temporisé de la théorie des codes sur canal contraint (v) énumérons et générons aléatoirement des permutations dans une certaine classe / Since early 90s, timed automata and timed languages are extensively used for modelling and verification of real-time systems, and thoroughly explored from a theoretical standpoint. Recently Asarin and Degorre introduced the notions of volume and entropy of timed languages to quantify the size of these languages and the information content of their elements. In this thesis we build new developments of this theory (called by us volumetry of timed languages) and apply it to several problems occurring in various domains of theoretical computer science such as verification, enumerative combinatorics or information theory. Among other we (i) develop a theory of timed symbolic dynamics; (ii) characterize a dichotomy between bad behaving and well behaving timed automata; (iii) define a least biased stochastic process for a timed automaton; (iv) develop a timed theory of constrained channel coding; (v)count and generate randomly and uniformly permutations in certain classes

Uma série de problemas de filas em sistemas de transportes com picos de tráfego, ou um número finito de elementos no sistema, são usualmente representados pelo modelo de fonte infinita, dadas as dificuldades de utilização do modelo exato de fonte finita. Este trabalho apresenta uma solução alternativa baseada no cálculo de um limitante superior para as probabilidades de equilíbrio do modelo exato de fonte finita, e compara as medidas de desempenho dos sistemas calculadas pelos dois modelos. Mostra-se que para índices de congestionamento menores que um, as diferenças entre estas medidas são tanto menores quanto menor for este índice. A partir destes resultados, compara-se as medidas de desempenho do modelo aproximado proposto com as do modelo de população infinita, para diferentes tamanhos de população e números de servidores. Conclui-se que os modelos conduzem a resultados numéricos muito próximos para uma ampla variação do índice de congestionamento, e que estes resultados são tão melhores quanto maior for o número de servidores no sistema e o número de elementos na população. São também apresentados três estudos de casos comparando os resultados induzidos pelos modelos exato, proposto e usual de fonte infinita, que ilustram a aplicabilidade prática dos resultados deste trabalho em sistemas de transportes. / A set of stochastic queueing problems in transportation systems with traffic peaks, or a finite number of elements in the system, are usually represented by the infinite source model, due to the difficulties of applying the exact finite source model. This study presents an alternative solution based on the upper bound values of the equilibrium probabilities of the exact finite source model, and compares the performance measurements of the two models. It is shown that for congestion factors below one, the smaller the value of the congestion factor, the smaller is the difference between the models. Based on this results the measures of performance of the proposed aproximate model are compared with the results of the usual infinite source model for different population sizes and number of servers. It is concluded that the models lead to very close numerical results for a wide range of congestion factors of the system and that these results are the better the larger is the number of servers and the number of elements in the system. Three case studies ilustrating the pratical applicability of the results of this study to transportation systems are also presented.

Customer lifetime value (CLV) is a forecasted expectation of the future value of a customer to the firm. There are two customer behavioral components of CLV that represent a particular modeling challenge: 1) how many transactions we expect from a customer in the future, and 2) how likely it is the customer remains active. Existing CLV models like the Pareto/NBD are valuable managerial tools because they are able to provide forward-looking estimates of transaction patterns and customer churn when the event of a customer leaving is unobservable, which is typical for most noncontractual goods and services. The CLV model literature has for the most part maintained its original assumption that the number of customer transactions follows a stable transaction process. Yet there are many categories of noncontractual goods and services where the stable transaction rate assumption is violated, particularly seasonal purchase patterns. CLV model estimates are further biased when there is an excess of customers with no repeat transactions. To address these modeling challenges, within this thesis I develop a generalized CLV modeling framework that combines three elements necessary to reduce bias in model estimates: 1) the incorporation of time-varying covariates to model data with transaction rates that change over time, 2) a zero-inflated model specification for customers with no repeat transactions, and 3) generalizes to different transaction process distributions to better fit diverse customer transaction patterns. This CLV modeling framework provides firms better estimates of the future activity of their customers, a critical CRM application.

Markov processes have been widely studied and used for modeling problems. A Markov process has two main components (i.e., an evolution law and an initial distribution). Markov processes are not suitable for modeling some problems, for example, the problem of predicting a trajectory with a known destination. Such a problem has three main components: an origin, an evolution law, and a destination. The conditionally Markov (CM) process is a powerful mathematical tool for generalizing the Markov process. One class of CM processes, called $CM_L$, fits the above components of trajectories with a destination. The CM process combines the Markov property and conditioning. The CM process has various classes that are more general and powerful than the Markov process, are useful for modeling various problems, and possess many Markov-like attractive properties. Reciprocal processes were introduced in connection to a problem in quantum mechanics and have been studied for years. But the existing viewpoint for studying reciprocal processes is not revealing and may lead to complicated results which are not necessarily easy to apply. We define and study various classes of Gaussian CM sequences, obtain their models and characterizations, study their relationships, demonstrate their applications, and provide general guidelines for applying Gaussian CM sequences. We develop various results about Gaussian CM sequences to provide a foundation and tools for general application of Gaussian CM sequences including trajectory modeling and prediction. We initiate the CM viewpoint to study reciprocal processes, demonstrate its significance, obtain simple and easy to apply results for Gaussian reciprocal sequences, and recommend studying reciprocal processes from the CM viewpoint. For example, we present a relationship between CM and reciprocal processes that provides a foundation for studying reciprocal processes from the CM viewpoint. Then, we obtain a model for nonsingular Gaussian reciprocal sequences with white dynamic noise, which is easy to apply. Also, this model is extended to the case of singular sequences and its application is demonstrated. A model for singular sequences has not been possible for years based on the existing viewpoint for studying reciprocal processes. This demonstrates the significance of studying reciprocal processes from the CM viewpoint.

We consider a longitudinal system in which transitions between the states are governed by a discrete time finite state space stochastic process X. Our aim, using aggregated sample survey data of the form typically collected by official statistical agencies, is to undertake model based inference for the underlying process X. We will develop inferential techniques for continuing sample surveys of two distinct types. First, longitudinal surveys in which the same individuals are sampled in each cycle of the survey. Second, cross-sectional surveys which sample the same population in successive cycles but with no attempt to track particular individuals from one cycle to the next. Some of the basic results have appeared in Davis et al (2001) and Davis et al (2002).¶ Longitudinal surveys provide data in the form of transition frequencies between the states of X. In Chapter Two we develop a method for modelling and estimating the one-step transition probabilities in the case where X is a non-homogeneous Markov chain and transition frequencies are observed at unit time intervals. However, due to their expense, longitudinal surveys are typically conducted at widely, and sometimes irregularly, spaced time points. That is, the observable frequencies pertain to multi-step transitions. Continuing to assume the Markov property for X, in Chapter Three, we show that these multi-step transition frequencies can be stochastically interpolated to provide accurate estimates of the one-step transition probabilities of the underlying process. These estimates for a unit time increment can be used to calculate estimates of expected future occupation time, conditional on an individual’s state at initial point of observation, in the different states of X.¶ For reasons of cost, most statistical collections run by official agencies are cross-sectional sample surveys. The data observed from an on-going survey of this type are marginal frequencies in the states of X at a sequence of time points. In Chapter Four we develop a model based technique for estimating the marginal probabilities of X using data of this form. Note that, in contrast to the longitudinal case, the Markov assumption does not simplify inference based on marginal frequencies. The marginal probability estimates enable estimation of future occupation times (in each of the states of X) for an individual of unspecified initial state. However, in the applications of the technique that we discuss (see Sections 4.4 and 4.5) the estimated occupation times will be conditional on both gender and initial age of individuals.¶ The longitudinal data envisaged in Chapter Two is that obtained from the surveillance of the same sample in each cycle of an on-going survey. In practice, to preserve data quality it is necessary to control respondent burden using sample rotation. This is usually achieved using a mechanism known as rotation group sampling. In Chapter Five we consider the particular form of rotation group sampling used by the Australian Bureau of Statistics in their Monthly Labour Force Survey (from which official estimates of labour force participation rates are produced). We show that our approach to estimating the one-step transition probabilities of X from transition frequencies observed at incremental time intervals, developed in Chapter Two, can be modified to deal with data collected under this sample rotation scheme. Furthermore, we show that valid inference is possible even when the Markov property does not hold for the underlying process.