Afin d'enrichir les données de corpus bilingues parallèles, il peut être judicieux de travailler avec des corpus dits comparables. En effet dans ce type de corpus, même si les documents dans la langue cible ne sont pas l'exacte traduction de ceux dans la langue source, on peut y retrouver des mots ou des phrases en relation de traduction. L'encyclopédie libre Wikipédia constitue un corpus comparable multilingue de plusieurs millions de documents. Notre travail consiste à trouver une méthode générale et endogène permettant d'extraire un maximum de phrases parallèles. Nous travaillons avec le couple de langues français-anglais mais notre méthode, qui n'utilise aucune ressource bilingue extérieure, peut s'appliquer à tout autre couple de langues. Elle se décompose en deux étapes. La première consiste à détecter les paires d’articles qui ont le plus de chance de contenir des traductions. Nous utilisons pour cela un réseau de neurones entraîné sur un petit ensemble de données constitué d'articles alignés au niveau des phrases. La deuxième étape effectue la sélection des paires de phrases grâce à un autre réseau de neurones dont les sorties sont alors réinterprétées par un algorithme d'optimisation combinatoire et une heuristique d'extension. L'ajout des quelques 560~000 paires de phrases extraites de Wikipédia au corpus d'entraînement d'un système de traduction automatique statistique de référence permet d'améliorer la qualité des traductions produites. Nous mettons les données alignées et le corpus extrait à la disposition de la communauté scientifique. / Working with comparable corpora can be useful to enhance bilingual parallel corpora. In fact, in such corpora, even if the documents in the target language are not the exact translation of those in the source language, one can still find translated words or sentences. The free encyclopedia Wikipedia is a multilingual comparable corpus of several millions of documents. Our task is to find a general endogenous method for extracting a maximum of parallel sentences from this source. We are working with the English-French language pair but our method -- which uses no external bilingual resources -- can be applied to any other language pair. It can best be described in two steps. The first one consists of detecting article pairs that are most likely to contain translations. This is achieved through a neural network trained on a small data set composed of sentence aligned articles. The second step is to perform the selection of sentence pairs through another neural network whose outputs are then re-interpreted by a combinatorial optimization algorithm and an extension heuristic. The addition of the 560~000 pairs of sentences extracted from Wikipedia to the training set of a baseline statistical machine translation system improves the quality of the resulting translations. We make both the aligned data and the extracted corpus available to the scientific community.

Bien que l’anesthésie générale soit un outil indispensable dans la chirurgie médicale d’aujourd’hui, ses mécanismes sous-jacents précis sont encore inconnus. Au cours de la sédation induite par le propofol les actions anesthésiques à l’échelle microscopique du neurone isolé conduisent à des changements spécifiques à l’échelle macroscopique qui sont observables comme les signaux électroencéphalogrammes (EEG). Pour une concentration faible en propofol, ces changements caractéristiques comprennent une augmentation de l’activité dans les bandes de fréquence delta (0.5-4 Hz) et alpha (8 13 Hz) dans la région frontal, une l’activité augmentée de delta et une l’activité diminuée de alpha dans la région occipitale. Dans cette thèse, nous utilisons des modèles de populations neuronales thalamo-corticales basés sur des données expérimentales. Les effets de propofol sur les synapses et sur les récepteurs extra-synaptiques GABAergiques situés dans le cortex et le thalamus sont modélisés afin de comprendre les mécanismes sous-jacents aux changements observés dans certaines puissances de l’EEG spectrale. Il est démontré que les modèles reproduisent bien les spectrales caractéristiques observées expérimentalement. Une des conclusions principales de ce travail est que l’origine des delta rythmes est fondamentalement différente de celle des alpha rythmes. Nos résultats indiquent qu’en fonction des valeurs moyennes des potentiels de l’état du système au repos, une augmentation ou une diminution des fonctions de gain thalamo-corticale résulte respectivement en une augmentation ou une diminution de alpha puissance. En revanche, l’évolution de la delta puissance est plutôt indépendant de l’état du système au repos; l'amélioration de la puissance spectrale de delta bande résulte de l’inhibition GABAergique synaptique ou extra-synaptique pour les fonctions de gain non linéaire à la fois croissante et décroissante. De plus, nous cherchons à identifier les paramètres d’un modèle de thalamo-corticale en ajustant le spectre de puissance de modèle pour les enregistrements EEG. Pour ce faire, nous considérons la tâche de l’estimation des paramètres dans les modèles qui sont décrits par un ensemble d’équations différentielles ordinaires ou bien stochastiques avec retard. Deux études de cas portant sur des données pseudo-expérimentales bruyantes sont d’abord effectuées pour comparer les performances des différentes méthodes d’optimisation. Les résultats de cette élaboration montrent que la méthode utilisée dans cette étude est capable d’estimer avec précision les paramètres indépendants du modèle et cela nous permet d’éviter les coûts de calcul des intégrations numériques. En considérant l’ensemble, les conclusions de cette thèse apportent de nouveaux éclairages sur les mécanismes responsables des changements spécifiques qui sont observées pendant la sédation propofol-induite dans les modèles de EEG. / Although general anaesthesia is an indispensable tool in today’s medical surgery, its precise underlying mechanisms are still unknown. During the propofol-induced sedation, the anaesthetic actions on the microscopic single neuron scale lead to specific changes in macroscopic-scale observables such as electroencephalogram (EEG) signals. For low concentration of propofol these characteristic changes comprised increased activity in the delta (0.5-4 Hz) and alpha (8-13 Hz) frequency bands over the frontal head region, but increased delta and decreased alpha power activity over the occipital region. In this thesis, we employ thalamo-cortical neural population models, and based on the experimental data, the propofol effects on the synaptic and extrasynaptic GABAergic receptors located in the cortex and thalamus are modelized to understand the mechanisms underlying the observed certain changes in EEG-spectral power. It is shown that the models reproduce well the characteristic spectral features observed experimentally. A key finding of this work is that the origin of delta rhythm is fundamentally different from the alpha rhythm. Our results indicate that dependent on the mean potential values of the system resting states, an increase or decrease in the thalamo-cortical gain functions results in an increase or decrease in the alpha power, respectively. In contrast, the evolution of the delta power is rather independent of the system resting states; the enhancement of spectral power in the delta band results from the increased synaptic or extra-synaptic GABAergic inhibition for both increasing and decreasing nonlinear gain functions. Furthermore, we aim to identify the parameters of a thalamo-cortical model by fitting the model power spectrum to the EEG recordings. To this end, we address the task of parameter estimation in the models that are described by a set of stochastic ordinary or delay differential equations. Two case studies dealing with noisy pseudo-experimental data are first carried out to compare the performance of different optimization methods. The results of this elaboration show that the method used in this study is able to accurately estimate the independent model parameters while it allows us to avoid the computational costs of the numerical integrations. Taken together, the findings of this thesis provide new insights into the mechanisms responsible for the specific changes in EEG patterns that are observed during propofol-induced sedation.

A satellite launch vehicle not subjected to any perturbations, external or internal, could be guided along a trajectory by following a stored, pre-computed steering program. In practice, perturbations do occur, and in order to take account of them and to achieve an accurate injection, a closed loop guidance system is required. Guidance algorithm is developed by solving the optimal control problem. Closed form solution is difficult because the necessary conditions are in the form of Two Point Boundary Value Problems (TBVP) or Multi Point Boundary Value Problems (MPBVP). Development of non-iterative guidance algorithm is taken as a prime objective of this thesis to ensure reliable on-board implementation. If non-iterative algorithms are required, the usual practice is to approximate the system equations to derive closed form solutions. In the present work, approximations cannot be used because the algorithm has to cater to a wide variety of vehicles and missions. Present development adopts an alternate approach by splitting the reconfigurable algorithm development in to smaller sub-problems such that each sub-problem has closed form solution. The splitting is done in such a way that the solution of the sub-problems can be used as building blocks to construct the final solution. By adding or removing the building blocks, the algorithm can be configured to suit specific requirements. Chapter 1 discusses the motivation and objectives of the thesis and gives a literature survey. In chapter 2, Classical Flat Earth (CFE) guidance algorithm is discussed. The assumptions and the nature of solution are closely analyzed because CFE guidance is used as the baseline for further developments. New contribution in chapter 2 is the extension of CFE guidance for a generalized propulsion system in which liquid and solid engines are present. In chapter 3, CFE guidance is applied for a mission with large pitch steering angles. The result shows loss of optimality and performance. An algorithm based on regular perturbation is developed to compensate for the small angle approximation. The new contribution in chapter 3 is the development of Regular Perturbation based FE (RPFE) guidance as an extension of CFE guidance. RPFE guidance can be configured as CFE guidance and FEGP. Algorithms presented up to chapter 3 are developed to inject a satellite in to orbits with unspecified inertial orientation. Communication satellite missions demand injection in to an orbit with a specific inertial orientation defined by argument of perigee. This problem is formulated using Calculus of Variations in chapter 4. A non-iterative closed form solution (Predicted target Flat Earth or PFE guidance) is derived for this problem. In chapter 5, PFE guidance is extended to a multi-stage vehicle with a constraint on the impact point of spent lower stage. Since the problem is not analytically solvable, the original problem is split in to three sub-problems and solved. Chapter 6 has two parts. First part gives theoretical analysis of the sub-optimal strategies with special emphasis to guidance. Behavior of predicted terminal error and control commands in presence of plant approximations are theoretically analyzed for a class of optimal control problems and the results are presented as six theorems. Chapter 7 presents the conclusions and future works.

Afin d'enrichir les données de corpus bilingues parallèles, il peut être judicieux de travailler avec des corpus dits comparables. En effet dans ce type de corpus, même si les documents dans la langue cible ne sont pas l'exacte traduction de ceux dans la langue source, on peut y retrouver des mots ou des phrases en relation de traduction. L'encyclopédie libre Wikipédia constitue un corpus comparable multilingue de plusieurs millions de documents. Notre travail consiste à trouver une méthode générale et endogène permettant d'extraire un maximum de phrases parallèles. Nous travaillons avec le couple de langues français-anglais mais notre méthode, qui n'utilise aucune ressource bilingue extérieure, peut s'appliquer à tout autre couple de langues. Elle se décompose en deux étapes. La première consiste à détecter les paires d’articles qui ont le plus de chance de contenir des traductions. Nous utilisons pour cela un réseau de neurones entraîné sur un petit ensemble de données constitué d'articles alignés au niveau des phrases. La deuxième étape effectue la sélection des paires de phrases grâce à un autre réseau de neurones dont les sorties sont alors réinterprétées par un algorithme d'optimisation combinatoire et une heuristique d'extension. L'ajout des quelques 560~000 paires de phrases extraites de Wikipédia au corpus d'entraînement d'un système de traduction automatique statistique de référence permet d'améliorer la qualité des traductions produites. Nous mettons les données alignées et le corpus extrait à la disposition de la communauté scientifique. / Working with comparable corpora can be useful to enhance bilingual parallel corpora. In fact, in such corpora, even if the documents in the target language are not the exact translation of those in the source language, one can still find translated words or sentences. The free encyclopedia Wikipedia is a multilingual comparable corpus of several millions of documents. Our task is to find a general endogenous method for extracting a maximum of parallel sentences from this source. We are working with the English-French language pair but our method -- which uses no external bilingual resources -- can be applied to any other language pair. It can best be described in two steps. The first one consists of detecting article pairs that are most likely to contain translations. This is achieved through a neural network trained on a small data set composed of sentence aligned articles. The second step is to perform the selection of sentence pairs through another neural network whose outputs are then re-interpreted by a combinatorial optimization algorithm and an extension heuristic. The addition of the 560~000 pairs of sentences extracted from Wikipedia to the training set of a baseline statistical machine translation system improves the quality of the resulting translations. We make both the aligned data and the extracted corpus available to the scientific community.

Optimization is a very important field with diverse applications in physical, social and biological sciences and in various areas of engineering. It appears widely in ma-chine learning, information retrieval, regression, estimation, operations research and a wide variety of computing domains. The subject is being deeply studied both theoretically and experimentally and several algorithms are available in the literature. These algorithms which can be executed (sequentially or concurrently) on a computing machine explore the space of input parameters to seek high quality solutions to the optimization problem with the search mostly guided by certain structural properties of the objective function. In certain situations, the setting might additionally demand for “absolute optimum” or solutions close to it, which makes the task even more challenging. In this thesis, we propose an optimization algorithm which is “gradient-free”, i.e., does not employ any knowledge of the gradient or higher order derivatives of the objective function, rather utilizes objective function values themselves to steer the search. The proposed algorithm is particularly effective in a black-box setting, where a closed-form expression of the objective function is unavailable and gradient or higher-order derivatives are hard to compute or estimate. Our algorithm is inspired by the well known cross entropy (CE) method. The CE method is a model based search method to solve continuous/discrete multi-extremal optimization problems, where the objective function has minimal structure. The proposed method seeks, in the statistical manifold of the parameters which identify the probability distribution/model defined over the input space to find the degenerate distribution concentrated on the global optima (assumed to be finite in quantity). In the early part of the thesis, we propose a novel stochastic approximation version of the CE method to the unconstrained optimization problem, where the objective function is real-valued and deterministic. The basis of the algorithm is a stochastic process of model parameters which is probabilistically dependent on the past history, where we reuse all the previous samples obtained in the process till the current instant based on discounted averaging. This approach can save the overall computational and storage cost. Our algorithm is incremental in nature and possesses attractive features such as stability, computational and storage efficiency and better accuracy. We further investigate, both theoretically and empirically, the asymptotic behaviour of the algorithm and find that the proposed algorithm exhibits global optimum convergence for a particular class of objective functions. Further, we extend the algorithm to solve the simulation/stochastic optimization problem. In stochastic optimization, the objective function possesses a stochastic characteristic, where the underlying probability distribution in most cases is hard to comprehend and quantify. This begets a more challenging optimization problem, where the ostentatious nature is primarily due to the hardness in computing the objective function values for various input parameters with absolute certainty. In this case, one can only hope to obtain noise corrupted objective function values for various input parameters. Settings of this kind can be found in scenarios where the objective function is evaluated using a continuously evolving dynamical system or through a simulation. We propose a multi-timescale stochastic approximation algorithm, where we integrate an additional timescale to accommodate the noisy measurements and decimate the effects of the gratuitous noise asymptotically. We found that if the objective function and the noise involved in the measurements are well behaved and the timescales are compatible, then our algorithm can generate high quality solutions. In the later part of the thesis, we propose algorithms for reinforcement learning/Markov decision processes using the optimization techniques we developed in the early stage. MDP can be considered as a generalized framework for modelling planning under uncertainty. We provide a novel algorithm for the problem of prediction in reinforcement learning, i.e., estimating the value function of a given stationary policy of a model free MDP (with large state and action spaces) using the linear function approximation architecture. Here, the value function is defined as the long-run average of the discounted transition costs. The resource requirement of the proposed method in terms of computational and storage cost scales quadratically in the size of the feature set. The algorithm is an adaptation of the multi-timescale variant of the CE method proposed in the earlier part of the thesis for simulation optimization. We also provide both theoretical and empirical evidence to corroborate the credibility and effectiveness of the approach. In the final part of the thesis, we consider a modified version of the control problem in a model free MDP with large state and action spaces. The control problem most commonly addressed in the literature is to find an optimal policy which maximizes the value function, i.e., the long-run average of the discounted transition payoffs. The contemporary methods also presume access to a generative model/simulator of the MDP with the hidden premise that observations of the system behaviour in the form of sample trajectories can be obtained with ease from the model. In this thesis, we consider a modified version, where the cost function to be optimized is a real-valued performance function (possibly non-convex) of the value function. Additionally, one has to seek the optimal policy without presuming access to the generative model. In this thesis, we propose a stochastic approximation algorithm for this peculiar control problem. The only information, we presuppose, available to the algorithm is the sample trajectory generated using a priori chosen behaviour policy. The algorithm is data (sample trajectory) efficient, stable, robust as well as computationally and storage efficient. We provide a proof of convergence of our algorithm to a high performing policy relative to the behaviour policy.

Made available in DSpace on 2014-12-17T14:54:53Z (GMT). No. of bitstreams: 1 RossanaMSC.pdf: 3237270 bytes, checksum: 01cfb4de4da5c1c94fba895ebbbdddb1 (MD5) Previous issue date: 2009-09-28 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / The bidimensional periodic structures called frequency selective surfaces have been well investigated because of their filtering properties. Similar to the filters that work at the traditional radiofrequency band, such structures can behave as band-stop or pass-band filters, depending on the elements of the array (patch or aperture, respectively) and can be used for a variety of applications, such as: radomes, dichroic reflectors, waveguide filters, artificial magnetic conductors, microwave absorbers etc. To provide high-performance filtering properties at microwave bands, electromagnetic engineers have investigated various types of periodic structures: reconfigurable frequency selective screens, multilayered selective filters, as well as periodic arrays printed on anisotropic dielectric substrates and composed by fractal elements. In general, there is no closed form solution directly from a given desired frequency response to a corresponding device; thus, the analysis of its scattering characteristics requires the application of rigorous full-wave techniques. Besides that, due to the computational complexity of using a full-wave simulator to evaluate the frequency selective surface scattering variables, many electromagnetic engineers still use trial-and-error process until to achieve a given design criterion. As this procedure is very laborious and human dependent, optimization techniques are required to design practical periodic structures with desired filter specifications. Some authors have been employed neural networks and natural optimization algorithms, such as the genetic algorithms and the particle swarm optimization for the frequency selective surface design and optimization. This work has as objective the accomplishment of a rigorous study about the electromagnetic behavior of the periodic structures, enabling the design of efficient devices applied to microwave band. For this, artificial neural networks are used together with natural optimization techniques, allowing the accurate and efficient investigation of various types of frequency selective surfaces, in a simple and fast manner, becoming a powerful tool for the design and optimization of such structures / As estruturas planares peri?dicas bidimensionais, conhecidas como Superf?cies Seletivas de Frequ?ncia, t?m sido bastante estudadas por causa da propriedade de filtragem de frequ?ncia que apresentam. Similares aos filtros que operam na faixa tradicional de radiofrequ?ncia, tais estruturas podem apresentar caracter?sticas espectrais de filtros rejeitafaixa ou passa-faixa, dependendo do tipo de elemento do arranjo (patch ou abertura, respectivamente) e podem ser utilizadas em uma variedade de aplica??es, tais como radomes, refletores dicr?icos, filtros de micro-ondas, condutores magn?ticos artificiais, absorvedores etc. Para melhorar o desempenho de tais dispositivos eletromagn?ticos e investigar suas propriedades, muitos estudiosos t?m analisado v?rios tipos de estruturas peri?dicas: superf?cies seletivas de frequ?ncia reconfigur?veis, filtros de m?ltiplas camadas seletivas, al?m de arranjos peri?dicos impressos sobre substratos diel?tricos anisotr?picos e que utilizam geometrias fractais na sua forma??o. Em geral, n?o existe uma solu??o anal?tica diretamente extra?da a partir da resposta em frequ?ncia de um dispositivo; desta forma, a an?lise de suas caracter?sticas espectrais requer a aplica??o de t?cnicas de onda completa rigorosas, como o m?todo da equa??o integral, por exemplo. Al?m disso, devido ? complexidade computacional exigida para a implementa??o destes m?todos, muitos estudiosos ainda utilizam a investiga??o por tentativa e erro, para alcan?ar crit?rios satisfat?rios ao projeto dos dispositivos. Como este procedimento ? muito trabalhoso e dependente do homem, faz-se necess?rio o emprego de t?cnicas de otimiza??o que acelerem a obten??o de estruturas peri?dicas com especifica??es de filtragem desejadas. Alguns autores t?m utilizado redes neurais artificiais e algoritmos de otimiza??o natural, como os algoritmos gen?ticos e a otimiza??o por enxame de part?culas no projeto e otimiza??o das superf?cies seletivas de frequ?ncia. Este trabalho tem como objetivo realizar um estudo mais aprofundado sobre o comportamento eletromagn?tico das estruturas peri?dicas seletivas de frequ?ncia, possibilitando a obten??o de dispositivos eficientes e aplic?veis na faixa de micro-ondas. P ra isto, redes neurais artificiais s?o utilizadas em conjunto com t?cnicas de otimiza??o baseadas na natureza, permitindo a investiga??o precisa e eficiente de v?rios tipos de superf?cies seletivas de frequ?ncia, de forma simples e r?pida, tornando-se, portanto, uma poderosa ferramenta de projeto e otimiza??o de tais estruturas

Submitted by Renata Lopes (renatasil82@gmail.com) on 2017-04-24T12:12:06Z No. of bitstreams: 1 eduardopestanadeaguiar.pdf: 1937921 bytes, checksum: 0472ffffb70cabf120dc5de86d6626b1 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2017-04-24T16:55:00Z (GMT) No. of bitstreams: 1 eduardopestanadeaguiar.pdf: 1937921 bytes, checksum: 0472ffffb70cabf120dc5de86d6626b1 (MD5) / Made available in DSpace on 2017-04-24T16:55:00Z (GMT). No. of bitstreams: 1 eduardopestanadeaguiar.pdf: 1937921 bytes, checksum: 0472ffffb70cabf120dc5de86d6626b1 (MD5) Previous issue date: 2011-08-30 / A presente dissertação tem como objetivo discutir o uso de técnicas de otimização baseadas no gradiente conjugado e de informações de segunda ordem para o treinamento de sistemas de inferência fuzzy singleton e non-singleton. Além disso, as soluções computacionais derivadas são aplicadas aos problemas de classificação de distúrbios múltiplos e isolados em sinais elétricos. Os resultados computacionais, obtidos a partir de dados sintéticos de distúrbios em sinais de tensão, indicam que os sistemas de inferência fuzzy singleton e non-singleton treinados pelos algoritmos de otimização considerados apresentam maior velocidade de convergência e melhores taxas de classificação quando comparados com aqueles treinados pelo algoritmo de otimização baseada em informações de primeira ordem e é bastante competitivo em relação à rede neural artificial perceptron multicamadas - multilayer perceptron (MLP) e ao classificador de Bayes. / This master dissertation aims to discuss the use of optimization techniques based on the conjugated gradient and on second order information for the training of singleton or non-singleton fuzzy inference systems. In addition, the computacional solutions obtained are applied to isolated a multiple disturbances classification problems in electric signals. Computational results obtained from synthetic data from disturbances in electric signals indicate that singleton or non-singleton fuzzy inference systems trained by the considered optimization algorithms present greater convergence speed and better classification rates when compared to those data trained by an optimization algorithm based on first order information and is quite competitive with multilayer perceptron neural network and Bayesian classifier.

This thesis is organized into two parts, one for my main area of research in the field of stochastic games, and the other for my contributions in the area of service systems. We first provide an abstract for my work in stochastic games. The field of stochastic games has been actively pursued over the last seven decades because of several of its important applications in oligopolistic economics. In the past, zero-sum stochastic games have been modelled and solved for Nash equilibria using the standard techniques of Markov decision processes. General-sum stochastic games on the contrary have posed difficulty as they cannot be reduced to Markov decision processes. Over the past few decades the quest for algorithms to compute Nash equilibria in general-sum stochastic games has intensified and several important algorithms such as stochastic tracing procedure [Herings and Peeters, 2004], NashQ [Hu and Wellman, 2003], FFQ [Littman, 2001], etc., and their generalised representations such as the optimization problem formulations for various reward structures [Filar and Vrieze, 1997] have been proposed. However, they suffer from either lack of generality or are intractable for even medium sized problems or both. In our venture towards algorithms for stochastic games, we start with a non-linear optimization problem and then design a simple gradient descent procedure for the same. Though this procedure gives the Nash equilibrium for a sample problem of terrain exploration, we observe that, in general, it need not be true. We characterize the necessary conditions and define KKT-N point. KKT-N points are those Karush-Kuhn-Tucker (KKT) points which corresponding to Nash equilibria. Thus, for a simple gradient based algorithm to guarantee convergence to Nash equilibrium, all KKT points of the optimization problem need to be KKT-N points, which restricts the applicability of such algorithms. We then take a step back and start looking at better characterization of those points of the optimization problem which correspond to Nash equilibria of the underlying game. As a result of this exploration, we derive two sets of necessary and sufficient conditions. The first set, KKT-SP conditions, is inspired from KKT conditions itself and is obtained by breaking down the main optimization problem into several sub-problems and then applying KKT conditions to each one of those sub-problems. The second set, SG-SP conditions, is a simplified set of conditions which characterize those Nash points more compactly. Using both KKT-SP and SG-SP conditions, we propose three algorithms, OFF-SGSP, ON-SGSP and DON-SGSP, respectively, which we show provide Nash equilibrium strategies for general-sum discounted stochastic games. Here OFF-SGSP is an off-line algorithm while ONSGSP and DON-SGSP are on-line algorithms. In particular, we believe that DON-SGSP is the first decentralized on-line algorithm for general-sum discounted stochastic games. We show that both our on-line algorithms are computationally efficient. In fact, we show that DON-SGSP is not only applicable for multi-agent scenarios but is also directly applicable for the single-agent case, i.e., MDPs (Markov Decision Processes). The second part of the thesis focuses on formulating and solving the problem of minimizing the labour-cost in service systems. We define the setting of service systems and then model the labour-cost problem as a constrained discrete parameter Markov-cost process. This Markov process is parametrized by the number of workers in various shifts and with various skill levels. With the number of workers as optimization variables, we provide a detailed formulation of a constrained optimization problem where the objective is the expected long-run averages of the single-stage labour-costs, and the main set of constraints are the expected long-run average of aggregate SLAs (Service Level Agreements). For this constrained optimization problem, we provide two stochastic optimization algorithms, SASOC-SF-N and SASOC-SF-C, which use smoothed functional approaches to estimate gradient and perform gradient descent in the aforementioned constrained optimization problem. SASOC-SF-N uses Gaussian distribution for smoothing while SASOC-SF-C uses Cauchy distribution for the same. SASOC-SF-C is the first Cauchy based smoothing algorithm which requires a fixed number (two) of simulations independent of the number of optimization variables. We show that these algorithms provide an order of magnitude better performance than existing industrial standard tool, OptQuest. We also show that SASOC-SF-C gives overall better performance.

In this thesis we apply neural networks as nonparametric and nonlinear methods to the Central European stock markets returns (Czech, Polish, Hungarian and German) modelling. In the first two chapters we define prediction task and link the classical econometric analysis to neural networks. We also present optimization methods which will be used in the tests, conjugate gradient, Levenberg-Marquardt, and evolutionary search method. Further on, we present statistical methods for comparing the predictive accuracy of the non-nested models, as well as economic significance measures. In the empirical tests we first show the power of neural networks on Mackey-Glass chaotic time series followed by real-world data of the daily and weekly returns of mentioned stock exchanges for the 2000:2006 period. We find neural networks to have significantly lower prediction error than classical models for daily DAX series, weekly PX50 and BUX series. The lags of time-series were used, and also cross-country predictability has been tested, but the results were not significantly different. We also achieved economic significance of predictions with both daily and weekly PX-50, BUX and DAX with 60% accuracy of prediction. Finally we use neural network to learn Black-Scholes model and compared the pricing errors of...

First part of the thesis is about literature research of optimization algorithms. Three of the algorithms were implemented and tested, concretely the ant colony algorithm, tabu search and simulated annealing. All three algorithms were implemented to solve the traveling salesman problem. In second part of the thesis the algorithms were tested and compared. In last part the influence of the ant colony parameters was evaluated.