Orientador: João Bosco Ribeiro do Val / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-06T12:35:26Z (GMT). No. of bitstreams: 1 Moises_GustavoViniciusLourenco_M.pdf: 6246087 bytes, checksum: c9946249f751bda598372a7979e3a189 (MD5) Previous issue date: 2005 / Resumo: Esta dissertação possui como tema a filtragem via Métodos de Monte Carlo para Cadeia de Markov. Através do estudo e da análise dos algoritmos de amostragem estocástica aliados às implementações numéricas, foi desenvolvida uma metodologia para avaliar e comparar as diversas técnicas de filtragem encontrados na literatura. Aplicações associando a filtragem recursiva ao controle via horizonte retrocedente também foram utilizadas para verificar o desempenho e a estabilidade do conjunto filtro/controle / Abstract: The dissertation's theme is the filtering problem via Markov Chain Monte Carlo methods. Combining the estudy and the analysis of the stochastic sampling algorithms with numerical implementations, we developted a methodology to evaluate and compare several filters in literature. Aplications of recursive filtering in association with receding horizon control tecniques were used to verify the finality and stability of the filter/control combination / Mestrado / Automação / Mestre em Engenharia Elétrica

Orientador: Jose Claudio Geromel / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-13T22:17:12Z (GMT). No. of bitstreams: 1 Goncalves_AlimPedrodeCastro_D.pdf: 1252972 bytes, checksum: 74f7eb667115d725da9b80029974a03f (MD5) Previous issue date: 2009 / Resumo: Este trabalho tem por objetivo o estudo do projeto de controladores dinâmicos H2 e H? por realimentação de saída para sistemas discretos com parâmetros sujeitos a saltos markovianos. Inicialmente, estudamos o caso de filtragem e propomos diferentes técnicas de projeto para casos especiais relacionados à disponibilidade do estado da cadeia de Markov, também chamado modo, ou a matriz de probabilidades de transição parcialmente conhecida. O resultado principal é a caracterização de todos os controladores dinâmicos lineares tais que a norma da saída controlada é limitada, levando a solução do problema de projeto do controlador linear ótimo H2 ou H?, sob a hipótese de conhecimento do modo por parte do controlador. Todos os projetos são expressos em termos de desigualdades matriciais lineares. Para ilustrar possíveis aplicações dos resultados teóricos, consideramos o projeto de um controlador cuja entrada é transmitida por um canal markoviano, além da modelagem estatística de falhas em sensores/atuadores / Abstract: This work addresses the H2 and H? output feedback design problem for discrete-time Markov jump linear systems. Initially, we study the filtering problem and propose different design techniques to deal with the Markov parameter, often called mode, availability and/or partly known transition probability. The main result is the characterization of all linear controllers such that the controlled output norm remains bounded by a given level, yielding the complete solution of the mode-dependent H2 and H? linear control design problem. All controllers are designed by solving linear matrix inequalities. The theory is illustrated by means of practical examples, consisting of control over data communication through a markovian channel and of statistical modelling of sensors/actuators failures / Doutorado / Automação / Doutor em Engenharia Elétrica

Orientadores: João Bosco Ribeiro do Val, Rafael Santos Mendes / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-16T00:17:25Z (GMT). No. of bitstreams: 1 Frencl_VictorBaptista_M.pdf: 1147363 bytes, checksum: 3933c461f1a19d86a46afeaba7057140 (MD5) Previous issue date: 2010 / Resumo: Esta dissertação possui como tema o estudo do problema de rastreamento de alvos manobrantes a partir da modelagem de sistemas dinâmicos com utilização da teoria de saltos markovianos nas transições entre modelos, da utilização de filtros estocásticos recursivos e de técnicas de filtragem. Foram feitos estudos e análises de dois tipos de modelos dinâmicos, o de velocidade constante e o de giro constante. Baseados nestes modelos, elaboraram-se algumas variações em cima destes. Também foram estudados modelos de observações, propondo a inclusão da velocidade radial nas observações do alvo. Os filtros estudados foram o filtro de Kalman estendido, que lida com modelos matemáticos não-lineares, e filtro BLUE, que trata de dinâmicas lineares e modelos de observações que envolvam conversões de coordenadas. As técnicas de filtragem de modelos múltiplos interagentes, que envolve chaveamento entre filtros, e de filtro de partículas, que baseia-se em simulações de Monte Carlo, foram estudados, propondo algumas variações destas técnicas. Foi desenvolvida uma metodologia, através de simulações numéricas no software MATLAB, para comparar desempenhos das propostas de técnicas de filtragem baseadas nestes estudos / Abstract: The dissertation's theme is the study of the maneuvering target tracking problem from dynamic systems modeling using markovian jumps on the transitions between models, recursive stochastic filters and filtering techniques. Surveys and analysis of two types of dynamic models were made: the constant velocity model and the constant turn model. Based on these models, some variations were prepared. Observations models were also studied, proposing the inclusion of the radial velocity in the target observations. The studied filters were the extended Kalman filter, which deals with nonlinear mathematical models, and the BLUE filter, which deals with linear dynamics and observations models which envolves coordinates conversions. The filtering techniques of the interacting multiple models, which involves the switching between models, and the particle filter, which is based on Monte Carlo simulations, were studied, proposing some variation of these techniques. We developed a methodology, using numerical simulations on MATLAB software, to compare performances of some of the filtering techniques based on these studies / Mestrado / Automação / Mestre em Engenharia Elétrica

Orientadores: Rubens Maciel Filho, Caliane Bastos Borba Costa / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Química / Made available in DSpace on 2018-08-16T03:24:44Z (GMT). No. of bitstreams: 1 Molano_LeonelMoreno_M.pdf: 2461312 bytes, checksum: 18c74e0ebd9c33e7cf133f1cb02b9064 (MD5) Previous issue date: 2010 / Resumo: A utilização de rotas verdes para obtenção de químicos é hoje uma área de grande interesse mundial. A necessidade de se desenvolver e estudar processos químicos que sejam, em sua integridade, independentes da via petroquímica são prementes. O etanal (acetaldeído), importante componente em diversos processos químicos, pode ser obtido pela oxidação do bioetanol. Entretanto, a conversão de bioetanol a acetaldeído é fortemente dependente da razão ar/etanol na alimentação e da temperatura. A baixas razões ar/etanol na alimentação, a conversão é mais baixa, mas a separação da corrente de produtos é menos custosa, em termos energéticos, pela inserção de menos nitrogênio no processo. Para obtenção de maiores conversões, altas razões de alimentação são requeridas, o que insere bastante nitrogênio ao processo, encarecendo a etapa de separação. Nesse sentido, este trabalho objetivou estudar a otimização do processo de obtenção de acetaldeído via oxidação do bioetanol em catalisador Fe-Mo. Para tanto, utilizou-se de modelagem já desenvolvida para o reator (em programação FORTRAN) e de simulador comercial para simular a etapa de separação da corrente de produtos. Visto que a modelagem do processo é não linear, algoritmos de otimização estocásticos foram utilizados na busca pela condição operacional. Os algoritmos estocásticos utilizados foram algoritmo genético, enxame de partícula e colônia de formigas, os quais foram usados e comparados, obtendo resultados das condições ótimas, tendo uma temperatura alimentação de 156,27°C , relação molar ar/etanol 16,60 e usando uma velocidade mássica de gás na alimentação de 3660 kg/m2h, requerendo uma energia especifica de 99,07 kJ/kg de acetaldeído para a separação dos produtos / Abstract Green routes used for the production of chemical products are nowadays of great interest worldwide. The need to develop and study chemical processes which are entirely independent on petrochemical route is urgent. Acetaldehyde is an important component in many chemical processes, and can be obtained by the oxidation of bioethanol. However, the conversion of bioethanol to acetaldehyde is highly dependent on the air/ethanol ratio and temperature in the reactor feed. At low air/ethanol ratios, the conversion is lower, but the separation of products is cheaper, in energetic terms, due to the lower insertion of nitrogen into the process. To obtain higher conversions, high feed ratios are required, but it turns more expensive the separation process because of the great amount of nitrogen in the process. The present work deals with the conceptual design of an optimized acetaldehyde production plant, through the oxidation of bioethanol at Fe-Mo catalysts. In order to perform the plant design and optimization, a detailed mathematical model is used to simulate the reactor, in FORTRAN language. To calculate the separation energy cost, designed with the commercial simulator. The optimization of operating conditions was made with Genetic Algorithm, Particle Swarm Algorithm and Ant Colony Optimization stochastic methods, due to the fact that the problem constraints are non-linear. In general, proves are search space for an optimal value function and used as operating conditions the temperature in the feeding of 156,27°C, Molar ratio Air / ethanol 16.6 0 and gas flow in the feeding of 3660 kg/m2h, requiring 99,07 kJ /kg of acetaldehyde to separate the products / Mestrado / Desenvolvimento de Processos Químicos / Mestre em Engenharia Química

Orientadores: Alim Pedro de Castro Gonçalves, Andre Ricardo Fioravanti / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-25T18:06:26Z (GMT). No. of bitstreams: 1 Cardeliquio_CaetanodeBrito_M.pdf: 2321696 bytes, checksum: b97dd016bf239aa4b647ee8c3411f3e1 (MD5) Previous issue date: 2014 / Resumo: Esta dissertação tem como objetivo principal o estudo de problemas de controle H2 e Hoo de sistemas lineares a tempo contínuo sujeitos a saltos markovianos com taxas de transição incertas. São tratados os problemas de realimentação de estado, realimentação de saída e filtragem. Todos os controladores e filtros são expressos em termos de desigualdades matriciais lineares / Abstract: This manuscript addresses the H2 and Hoo control for continuous time systems subjected to markovian jumps with uncertain transition rates. Our purpose is to minimize the H2 and Hoo norms using Linear Matrix Inequalities (LMIs) and find the gains for the state-feedback control when the transition rate between modes has some degree of uncertainty. Output feedback and filtering problems are both also addressed / Mestrado / Automação / Mestre em Engenharia Elétrica

No description available.

Orientador: João Bosco Ribeiro do Val / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-04T02:16:31Z (GMT). No. of bitstreams: 1 Oliveira_CristianeNespolide_D.pdf: 558682 bytes, checksum: bdb0231f0c5d9c4cc09a4232a2c21df8 (MD5) Previous issue date: 2004 / Resumo: Este trabalho aborda conceitos de estabilidade de segundo momento e um problema de controle ótimo envolvendo sistemas lineares com saltos Markovianos a tempo discreto. No modelo estudado, define-se como ho-rizonte um tempo de parada 't que representa a ocorrência de um número fixo N de falhas ou reparos (TN), ou a ocorrência de uma falha crucial ('t~), após as quais o sistema é paralisado para manutenção. Considera-se ainda, um caso misto intermediário onde 't representa o mínimo entre TN e 't~ . Estes tempos de parada coincidem com algum dos instantes de salto da cadeia de Markov e a informação disponível permite a reconfiguração da ação de controle em cada um destes instantes, na forma de um ganho de realimentação linear. Através do conceito denominado 't-estabilidade estocástica são obtidas condições necessárias e suficientes para a esta-bilidade estocástica do sistema até a ocorrência do tempo de parada 't~. Estas condições conduzem a um teste que se beneficia da estrutura da ca-deia para propor uma decomposição para a verificação de estabilidade em média quadrática, o que neste sentido, induz métodos algorítmicos mais simples. Adcionalmente, a equivalência entre conceitos de 't-estabilidade estocástica de segundo momento é estabelecida para os três tempos de pa-rada discriminados. A solução ótima para o problema de controle linear quadrático (LQ) por realimentação de estado, com ou sem ruído Gaussi-ano, tendo como horizonte os tempos de parada TN ou 't~ é apresentada. Esta solução é dada em termos de um conjunto de equações algébricas de Riccati recursivas ou um conjunto de equaç(les algébricas de Riccati aco-pIadas (EARA). Além disso, aborda-se o problema de controle LQG por realimentação dinâmica de saída via estimação de estado / Abstract: This work deals with second order stability concepts and a stochastic optimal control problem involving discrete-time jump Markov linear sys-tems. The jumps or changes between the system operation modes evolve according to an underlying Markov chain. In the mo deI studied, the pro-blem horizon is defined by a stopping time 't which represents either, the occurrence of a fixed number N of failures or repairs (TN), or the occur-rence of a crucial failure event ('td), after which the system is brought to a halt for maintenance. In addition, an intermediary mixed case for which 't represents the minimum between TN and 'td is also considered. These stopping times coincide with some of the jump times of the Markov state and the information available allows the reconfiguration of the control ac-tion at each jump time, in the form of a linear feedback gain. Using the concept named stochastic 't-stability, equivalent conditions to ensure the stochastic stability of the system until the occurrence of the stopping time 'td is obtained. These conditions lead to a test that benefits from the chain structure for proposing a simpler decomposition algorithm for the mean square stability verification. The work also develops equivalences among second order 't-stability concepts, for alI stopping times considered, that parallels the results for infinite horizon problems. Considering TN and 'td as horizon, the optimal control solution for the linear quadratic (LQ) pro-blem for state feedback, with or without Gaussian noise, is presented. The solution is given in terms of recursions of a set of algebraic Riccati equa-tions or a coupled set of algebraic Riccati equation (CARE). The LQG optimal control problem for dynamic output feedback using state estima-tion is also studied. / Doutorado / Automação / Doutor em Engenharia Elétrica

Cette thèse examine certaines questions quantitatives dans le cadre de deux modèles stochastiques différents. Il est divisé en deux parties : la première partie examine une nouvelle classe de jeux stochastiques avec une fonction de paiement particulière que nous appelons « de priorité ». Cette classe de jeux contient comme sous-classes propre les jeux de parité, largement étudiés en informatique, et les jeux de limsup et liminf, étudiés dans la théorie des jeux. La deuxième partie de la thèse examine certaines questions naturelles mais complexes sur les distributions, étudiées dans le cadre plus simple des chaînes de Markov à espace d'états fini. Dans la première partie, nous examinons les jeux à somme nulle à deux joueurs en se centrant sur la fonction de paiement de priorité. Cette fonction de paiement génère le gain utilisé dans les jeux de parité. Nous considérons à la fois les jeux de priorité stochastiques à tour de rôle et les jeux de priorité simultanés. Notre approche des jeux de priorité est basée sur le concept du point fixe le plus proche (« nearest fixed point ») des applications monotones non expansives et étend l'approche mu-calcul aux jeux de priorité.La deuxième partie de la thèse concerne les questions de population. De manière simplifiée, nous examinons comment une distribution de probabilité sur les états évolue dans le temps. Plus précisément, nous sommes intéressés par des questions comme la suivante : à partir d'une distribution initiale, la population peut-elle atteindre à un moment donné une distribution avec une probabilité dépassant un seuil donné dans l'état visé? Il s'avère que ce type de questions est beaucoup plus difficile à gérer que les questions concernant les trajectoires individuelles : on ne connaît pas, pour le modèle des chaînes de Markov, si les questions de population soient décidables. Nous étudions les restrictions des chaînes de Markov assurant la décision des questions de population. / This thesis examines some quantitative questions in the framework of two different stochastic models. It is divided into two parts: the first part examines a new class of stochastic games with priority payoff. This class of games contains as proper subclasses the parity games extensively studied in computer science, and limsup and liminf games studied in game theory. The second part of the thesis examines some natural but involved questions about distributions, studied in the simple framework of finite state Markov chain.In the first part, we examine two-player zero-sum games focusing on a particular payoff function that we call the priority payoff. This payoff function generalizes the payoff used in parity games. We consider both turn-based stochastic priority games and concurrent priority games. Our approach to priority games is based on the concept of the nearest fixed point of monotone nonexpansive mappings and extends the mu-calculus approach to priority games.The second part of the thesis deals with population questions. Roughly speaking, we examine how a probability distribution over states evolves in time. More specifically, we are interested in questions like the following one: from an initial distribution, can the population reach at some moment a distribution with a probability mass exceeding a given threshold in state Goal? It turns out that this type of questions is much more difficult to handle than the questions concerning individual trajectories: it is not known for the simple model of Markov chains whether population questions are decidable. We study restrictions of Markov chains ensuring decidability of population questions.

Machine learning is a major source of interesting optimization problems of current interest. These problems tend to be challenging because of their enormous scale, which makes it difficult to apply traditional optimization algorithms. We explore three avenues to designing algorithms suited to handling these challenges, with a view toward large-scale ML tasks. The first is to develop better general methods for unconstrained minimization. The second is to tailor methods to the features of modern systems, namely the availability of distributed computing. The third is to use specialized algorithms to exploit specific problem structure. Chapters 2 and 3 focus on improving quasi-Newton methods, a mainstay of unconstrained optimization. In Chapter 2, we analyze an extension of quasi-Newton methods wherein we use block updates, which add curvature information to the Hessian approximation on a higher-dimensional subspace. This defines a family of methods, Block BFGS, that form a spectrum between the classical BFGS method and Newton's method, in terms of the amount of curvature information used. We show that by adding a correction step, the Block BFGS method inherits the convergence guarantees of BFGS for deterministic problems, most notably a Q-superlinear convergence rate for strongly convex problems. To explore the tradeoff between reduced iterations and greater work per iteration of block methods, we present a set of numerical experiments. In Chapter 3, we focus on the problem of step size determination. To obviate the need for line searches, and for pre-computing fixed step sizes, we derive an analytic step size, which we call curvature-adaptive, for self-concordant functions. This adaptive step size allows us to generalize the damped Newton method of Nesterov to other iterative methods, including gradient descent and quasi-Newton methods. We provide simple proofs of convergence, including superlinear convergence for adaptive BFGS, allowing us to obtain superlinear convergence without line searches. In Chapter 4, we move from general algorithms to hardware-influenced algorithms. We consider a form of distributed stochastic gradient descent that we call Leader SGD, which is inspired by the Elastic Averaging SGD method. These methods are intended for distributed settings where communication between machines may be expensive, making it important to set their consensus mechanism. We show that LSGD avoids an issue with spurious stationary points that affects EASGD, and provide a convergence analysis of LSGD. In the stochastic strongly convex setting, LSGD converges at the rate O(1/k) with diminishing step sizes, matching other distributed methods. We also analyze the impact of varying communication delays, stochasticity in the selection of the leader points, and under what conditions LSGD may produce better search directions than the gradient alone. In Chapter 5, we switch again to focus on algorithms to exploit problem structure. Specifically, we consider problems where variables satisfy multiaffine constraints, which motivates us to apply the Alternating Direction Method of Multipliers (ADMM). Problems that can be formulated with such a structure include representation learning (e.g with dictionaries) and deep learning. We show that ADMM can be applied directly to multiaffine problems. By extending the theory of nonconvex ADMM, we prove that ADMM is convergent on multiaffine problems satisfying certain assumptions, and more broadly, analyze the theoretical properties of ADMM for general problems, investigating the effect of different types of structure.

The scheduling of material requirements at a factory to maximize profits.or productivity is a difficult mathematical problem. The stochastic nature of most production setups introduces additional complications as a result of the uncertainty involved in vendor reliability and processing times. But in developing the descriptive model for a system, a true representation can only be attained if the variability of these elements is considered. Here we present the development of a normative model based on a new type of descriptive model which considers the element of stochasticity. The arrival time of an assembly part from a vendor is considered to be a normally distributed random variable. We attempt to optimize the system with regard to work-in-process inventory using a dynamic programming algorithm in combination with a heuristic procedure. The decision variable is the prescribed assembly part delivery date. The model is particularly suitable for application in low volume assembly lines, where products are manufactured in discrete batches. / Master of Science