The estimation and analysis of signals that have polynomial phase and constant or time-varying amplitudes with the addititve noise is considered in this dissertation.Much work has been undertaken on this problem over the last decade or so, and there are a number of estimation schemes available. The fundamental problem when trying to estimate the parameters of these type of signals is the nonlinear characterstics of the signal, which lead to computationally difficulties when applying standard techniques such as maximum likelihood and least squares. When considering only the phase data, we also encounter the well known problem of the unobservability of the true noise phase curve. The methods that are currently most popular involve differencing in phase followed by regression, or nonlinear transformations. Although these methods perform quite well at high signal to noise ratios, their performance worsens at low signal to noise, and there may be significant bias. One of the biggest problems to efficient estimation of these models is that the majority of methods rely on sequential estimation of the phase coefficients, in that the highest-order parameter is estimated first, its contribution removed via demodulation, and the same procedure applied to estimation of the next parameter and so on. This is clearly an issue in that errors in estimation of high order parameters affect the ability to estimate the lower order parameters correctly. As a result, stastical analysis of the parameters is also difficult. In thie dissertation, we aim to circumvent the issues of bias and sequential estiamtion by considering the issue of full parameter iterative refinement techniques. ie. given a possibly biased initial estimate of the phase coefficients, we aim to create computationally efficient iterative refinement techniques to produce stastically efficient estimators at low signal to noise ratios. Updating will be done in a multivariable manner to remove inaccuracies and biases due to sequential procedures. Stastical analysis and extensive simulations attest to the performance of the schemes that are presented, which include likelihood, least squares and bayesian estimation schemes. Other results of importance to the full estimatin problem, namely when there is error in the time variable, the amplitude is not constant, and when the model order is not known, are also condsidered.

Modern financial asset pricing theory is a broad, and at times, complex field. The literature review in this study covers many of the asset pricing techniques including factor models, random walk models, correlation models, Bayesian methods, autoregressive models, moment-matching models, stochastic jumps and mean reversion models. An important topic in finance is portfolio opti-misation with respect to risk and reward such as the mean variance optimisation introduced by Markowitz (1952). This study covers optimisation techniques such as single period mean variance optimisation, optimisation with risk aversion, multi-period stochastic programs, two-fund separa- tion theory, downside optimisation techniques and multi-period optimisation such as the Bellman dynamic programming model. The question asked in this study is, in the context of investing for South African individuals in a multi-asset portfolio, whether an active investment strategy is signi cantly di erent from a passive investment strategy. The passive strategy is built using stochastic programming with moment matching methods for non-Gaussian asset class distributions. The strategy is optimised in a framework using a downside risk metric, the conditional variance at risk. The active strategy is built with forward forecasts for asset classes using the time-varying transitional-probability Markov regime switching model. The active portfolio is finalised by a dynamic optimisation using a two-stage stochastic programme with recourse, which is solved as a large linear program. A hypothesis test is used to establish whether the results of two strategies are statistically different. The performance of the strategies are also reviewed relative to multi-asset peer rankings. Lastly, we consider whether the findings reveal information on the degree of effi ciency in the market place for multi-asset investments for the South African investor. / Operations Management / M. Sc. (Operations Research)

Les réseaux dynamiques sont constitués d’entités établissant des contacts les unes avec les autres dans le temps. Un défi majeur dans les réseaux dynamiques est de prédire les modèles de mobilité et de décider si l’évolution de la topologie satisfait aux exigences du succès d’un algorithme donné. Les types de dynamique résultant de ces réseaux sont variés en échelle et en nature. Par exemple,certains de ces réseaux restent connexes tout le temps; d’autres sont toujours déconnectés mais offrent toujours une sorte de connexité dans le temps et dans l’espace(connexité temporelle); d’autres sont connexes de manière récurrente, périodique,etc. Tous ces contextes peuvent être représentés sous forme de classes de graphes dynamiques correspondant à des conditions nécessaires et/ou suffisantes pour des problèmes ou algorithmes distribués donnés. Étant donné un graphe dynamique,une question naturelle est de savoir à quelles classes appartient ce graphe. Dans ce travail, nous apportons une contribution à l’automatisation de la classification de graphes dynamiques. Nous proposons des stratégies pour tester l’appartenance d’un graphe dynamique à une classe donnée et nous définissons un cadre générique pour le test de propriétés dans les graphes dynamiques. Nous explorons également le cas où aucune propriété sur le graphe n’est garantie, à travers l’étude du problème de maintien d’une forêt d’arbres couvrants dans un graphe dynamique. / Dynamic networks consist of entities making contact over time with one another. A major challenge in dynamic networks is to predict mobility patterns and decide whether the evolution of the topology satisfies requirements for the successof a given algorithm. The types of dynamics resulting from these networks are varied in scale and nature. For instance, some of these networks remain connected at all times; others are always disconnected but still offer some kind of connectivity over time and space (temporal connectivity); others are recurrently connected,periodic, etc. All of these contexts can be represented as dynamic graph classes corresponding to necessary or sufficient conditions for given distributed problems or algorithms. Given a dynamic graph, a natural question to ask is to which of the classes this graph belongs. In this work we provide a contribution to the automation of dynamic graphs classification. We provide strategies for testing membership of a dynamic graph to a given class and a generic framework to test properties in dynamic graphs. We also attempt to understand what can still be done in a context where no property on the graph is guaranteed through the distributed problem of maintaining a spanning forest in highly dynamic graphs.

Esta Dissertação irá apresentar a utilização de técnicas de controle nãolinear, tais como o controle adaptativo e robusto, de modo a controlar um sistema de Eletroestimulação Funcional desenvolvido pelo laboratório de Engenharia Biomédica da COPPE/UFRJ. Basicamente um Eletroestimulador Funcional (Functional Electrical Stimulation FES) se baseia na estimulação dos nervos motores via eletrodos cutâneos de modo a movimentar (contrair ou distender) os músculos, visando o fortalecimento muscular, a ativação de vias nervosas (reinervação), manutenção da amplitude de movimento, controle de espasticidade muscular, retardo de atrofias e manutenção de tonicidade muscular. O sistema utilizado tem por objetivo movimentar os membros superiores através do estímulo elétrico de modo a atingir ângulos-alvo pré-determinados para a articulação do cotovelo. Devido ao fato de não termos conhecimento pleno do funcionamento neuro-motor humano e do mesmo ser variante no tempo, não-linear, com parâmetros incertos, sujeito a perturbações e completamente diferente para cada indivíduo, se faz necessário o uso de técnicas de controle avançadas na tentativa de se estabilizar e controlar esse tipo de sistema. O objetivo principal é verificar experimentalmente a eficácia dessas técnicas de controle não-linear e adaptativo em comparação às técnicas clássicas, de modo a alcançar um controle mais rápido, robusto e que tenha um desempenho satisfatório. Em face disso, espera-se ampliar o campo de utilização de técnicas de controle adaptativo e robusto, além de outras técnicas de sistemas inteligentes, tais como os algoritmos genéticos, provando que sua aplicação pode ser efetiva no campo de sistemas biológicos e biomédicos, auxiliando assim na melhoria do tratamento de pacientes envolvidos nas pesquisas desenvolvidas no Laboratório de Engenharia Biomédica da COPPE/UFRJ. / This dissertation will present the use of nonlinear control techniques, such as adaptive and robust control in order to design a Functional Electrical Stimulation (FES) system developed by Biomedical Engineering Laboratory at COPPE/UFRJ. Basically, a FES on the stimulation of motor nerves via skin electrodes in order to contract or stretch the muscles such that the amplitude and quality of the limbs movement can be maintained, reducing muscular atrophy as well. Consequently, the muscle strength can be improved and new neural pathways may be activated. Here, the goals of the proposed control system is to move the arm of the patient via electrical stimulation to achieve some desired trajectory related to the elbow angles of reference. Since we have a priori no deep knowledge of human neuro-motor model, the use of advanced and robust control schemes seems to be useful to stabilize this kind of systems which may be completely different for each individual, being time-varying, nonlinear, uncertain and subject to disturbances. The main objective is to experimentally verify the effectiveness of the proposed nonlinear and adaptive controllers when compared to classical ones in order to achieve faster, robust and better control performance. It is expected to spread the application of adaptive and robust controllers and other intelligent system tools, such as genetic algorithms, to the field of biological and biomedical engineering. Thus, we believe that the developed control system may help the improvement of the patients treatment involved in the research carried out by Biomedical Engineering Laboratory at COPPE/UFRJ.

Esta Dissertação irá apresentar a utilização de técnicas de controle nãolinear, tais como o controle adaptativo e robusto, de modo a controlar um sistema de Eletroestimulação Funcional desenvolvido pelo laboratório de Engenharia Biomédica da COPPE/UFRJ. Basicamente um Eletroestimulador Funcional (Functional Electrical Stimulation FES) se baseia na estimulação dos nervos motores via eletrodos cutâneos de modo a movimentar (contrair ou distender) os músculos, visando o fortalecimento muscular, a ativação de vias nervosas (reinervação), manutenção da amplitude de movimento, controle de espasticidade muscular, retardo de atrofias e manutenção de tonicidade muscular. O sistema utilizado tem por objetivo movimentar os membros superiores através do estímulo elétrico de modo a atingir ângulos-alvo pré-determinados para a articulação do cotovelo. Devido ao fato de não termos conhecimento pleno do funcionamento neuro-motor humano e do mesmo ser variante no tempo, não-linear, com parâmetros incertos, sujeito a perturbações e completamente diferente para cada indivíduo, se faz necessário o uso de técnicas de controle avançadas na tentativa de se estabilizar e controlar esse tipo de sistema. O objetivo principal é verificar experimentalmente a eficácia dessas técnicas de controle não-linear e adaptativo em comparação às técnicas clássicas, de modo a alcançar um controle mais rápido, robusto e que tenha um desempenho satisfatório. Em face disso, espera-se ampliar o campo de utilização de técnicas de controle adaptativo e robusto, além de outras técnicas de sistemas inteligentes, tais como os algoritmos genéticos, provando que sua aplicação pode ser efetiva no campo de sistemas biológicos e biomédicos, auxiliando assim na melhoria do tratamento de pacientes envolvidos nas pesquisas desenvolvidas no Laboratório de Engenharia Biomédica da COPPE/UFRJ. / This dissertation will present the use of nonlinear control techniques, such as adaptive and robust control in order to design a Functional Electrical Stimulation (FES) system developed by Biomedical Engineering Laboratory at COPPE/UFRJ. Basically, a FES on the stimulation of motor nerves via skin electrodes in order to contract or stretch the muscles such that the amplitude and quality of the limbs movement can be maintained, reducing muscular atrophy as well. Consequently, the muscle strength can be improved and new neural pathways may be activated. Here, the goals of the proposed control system is to move the arm of the patient via electrical stimulation to achieve some desired trajectory related to the elbow angles of reference. Since we have a priori no deep knowledge of human neuro-motor model, the use of advanced and robust control schemes seems to be useful to stabilize this kind of systems which may be completely different for each individual, being time-varying, nonlinear, uncertain and subject to disturbances. The main objective is to experimentally verify the effectiveness of the proposed nonlinear and adaptive controllers when compared to classical ones in order to achieve faster, robust and better control performance. It is expected to spread the application of adaptive and robust controllers and other intelligent system tools, such as genetic algorithms, to the field of biological and biomedical engineering. Thus, we believe that the developed control system may help the improvement of the patients treatment involved in the research carried out by Biomedical Engineering Laboratory at COPPE/UFRJ.

Guidance is defined as the determination of a strategy for following a nominal path in the presence of o-nominal conditions, disturbances and uncertainties, and the strategy employed is called a guidance law. Variants of Proportional Navigation (PN), such as True Proportional Navigation (TPN) and Pure Proportional Navigation (PPN), have been studied extensively in the literature on tactical missile guidance. In the absence of target maneuvers, in a linear interceptor guidance problem, TPN was shown to be optimal. However, the standard PN class of guidance laws per se does not show good performance against maneuvering targets, and was found to be eective in intercepting a maneuvering target only from a restrictive set of initial geometries. Also, since these guidance laws were eectively designed for lower speed targets, they show a degraded performance when applied against higher speed targets. However, in the current defense scenario, two classes of agile targets, which are capable of continuous maneuver, and/or of much higher speed than the interceptor, are a reality. This thesis presents analysis of several variants of PN class of guidance laws against these two classes of agile targets. In the literature, an augmentation of the TPN guidance law, termed as Augmented Proportional Navigation (APN), was shown to be optimal in linearized engagement framework. The present work proposes an augmentation of the PPN guidance law, which is more realistic than TPN for an aerodynamically controlled interceptor, and an-alyzes its capturability in fully nonlinear framework, and develops sauciest conditions on speed ratio, navigation gain and augmentation parameter to ensure that all possible initial engagement geometries are included in the capture zone when applied against a target executing piecewise continuous maneuver. The thesis also obtains the capture zone in the relative velocity space for augmented PPN guidance law. In the literature, a novel guidance law was proposed for the interception of higher speed targets in planar engagement by using a negative navigation gain instead of the standard positive one, and was termed as Retro-PN. It was shown that even though the Retro-PN guided interceptor takes more time than PN guided one in achieving successful interception, Retro-PN performs significantly better than the classical PN law, in terms of capturability, lateral acceleration demand, and closing velocity, when used against higher speed targets. The thesis analyzes Retro-PN guidance law in 3-D engagement geometries to yield the complete capture zone of interceptors guided by Retro-PN guidance philosophy, and derives necessary and sucient conditions for the capture of higher speed non-maneuvering targets with and without a constraint on finiteness of lateral acceleration. Terminal impact angle control is crucial for enhancement of warhead eectiveness. In the literature, this problem has been addressed mostly in the context of targets with lower speeds than the interceptor. The thesis analyzes the performance of a composite PN guidance law, that uses standard PPN and the Retro-PN guidance laws based on initial engagement geometry and requirement of impact angle, against higher speed non-maneuvering targets. Then, to expand the set of achievable impact angles, it proposes a modified composite PN guidance scheme, and analyzes the same. For implementation of many modern guidance laws, a good estimate of time-to-go is essential. This requirement is especially severe in case of impact time constrained en-gagement scenarios. To this end, an ecient and fast time-to-go estimation algorithm for generic 3-D engagement is required. Two time-to-go estimation algorithms are presented and analyzed in this work for the engagement of a PPN or Retro-PN guided interceptor and a higher speed target. The first one is a closed form approximation of time-to-go in terms of range, nominal closing speed and an indicator of heading error, and the second one is a numerical recursive time-to-go estimation algorithm. To improve the odds of intercepting an intelligent target and destroying it, a salvo attack of two or more interceptors could be considered as a viable option. Moreover, this simultaneous salvo attack can also be further improved in eciency by incorporating the shoot-look-shoot approach in making a decision about launching interceptors. This can be considered as the first step towards a layered defense system, which has been described in the literature as a potentially eective strategy against short range or long range ballistic threat. To this end, the present work proposes two PPN and Retro-PN based guidance strategies for achieving simultaneous salvo attack on a higher speed non-maneuvering target. For the implementation of the same the numerical recursive time-to-go estimation technique proposed in this work is utilized

Floods cause widespread damage to property and life in different parts of the world. Hence there is a paramount need to develop effective methods for design flood estimation to alleviate risk associated with these extreme hydrologic events. Methods that are conventionally considered for analysis of floods focus on estimation of continuous frequency relationship between peak flow observed at a location and its corresponding exceedance probability depicting the plausible conditions in the planning horizon. These methods are commonly known as at-site flood frequency analysis (FFA) procedures. The available FFA procedures can be classified as parametric and nonparametric. Parametric methods are based on the assumption that sample (at-site data) is drawn from a population with known probability density function (PDF). Those procedures have uncertainty associated with the choice of PDF and the method for estimation of its parameters. Moreover, parametric methods are ineffective in modeling flood data if multimodality is evident in their PDF. To overcome those artifacts, a few studies attempted using kernel based nonparametric (NP) methods as an alternative to parametric methods. The NP methods are data driven and they can characterize the uncertainty in data without prior assumptions as to the form of the PDF. Conventional kernel methods have shortcomings associated with boundary leakage problem and normal reference rule (considered for estimation of bandwidth), which have implications on flood quantile estimates. To alleviate this problem, focus of NP flood frequency analysis has been on development of new kernel density estimators (kdes). Another issue in FFA is that information on the whole hydrograph (e.g., time to the peak flow, volume of the flood flow and duration of the flood event) is needed, in addition to peak flow for certain applications. An option is to perform frequency analysis on each of the variables independently. However, these variables are not independent, and hence there is a need to perform multivariate analysis to construct multivariate PDFs and use the corresponding cumulative distribution functions (CDFs) to arrive at estimates of characteristics of design flood hydrograph. In this perspective, recent focus of flood frequency analysis studies has been on development of methods to derive joint distributions of flood hydrograph related variables in a nonparametric setting. Further, in real world scenario, it is often necessary to estimate design flood quantiles at target locations that have limited or no data. Regional Flood Frequency analysis (RFFA) procedures have been developed for use in such situations. These procedures involve use of a regionalization procedure for identification of a homogeneous group of watersheds that are similar to watershed of the target site in terms of flood response. Subsequently regional frequency analysis (RFA) is performed, wherein the information pooled from the group (region) forms basis for frequency analysis to construct a CDF (growth curve) that is subsequently used to arrive at quantile estimates at the target site. Though there are various procedures for RFFA, they are largely confined to only univariate framework considering a parametric approach as the basis to arrive at required quantile estimates. Motivated by these findings, this thesis concerns development of a linear diffusion process based adaptive kernel density estimator (D-kde) based methodologies for at-site as well as regional FFA in univariate as well as bivariate settings. The D-kde alleviates boundary leakage problem and also avoids normal reference rule while estimating optimal bandwidth by using Botev-Grotowski-Kroese estimator (BGKE). Potential of the proposed methodologies in both univariate and bivariate settings is demonstrated by application to synthetic data sets of various sizes drawn from known unimodal and bimodal parametric populations, and to real world data sets from India, USA, United Kingdom and Canada. In the context of at-site univariate FFA (considering peak flows), the performance of D- kde was found to be better when compared to four parametric distribution based methods (Generalized extreme value, Generalized logistic, Generalized Pareto, Generalized Normal), thirty-two ‘kde and bandwidth estimator’ combinations that resulted from application of four commonly used kernels in conjunction with eight bandwidth estimators, and a local polynomial–based estimator. In the context of at-site bivariate FFA considering ‘peakflow-flood volume’ and ‘flood duration-flood volume’ bivariate combinations, the proposed D-kde based methodology was shown to be effective when compared to commonly used seven copulas (Gumbel-Hougaard, Frank, Clayton, Joe, Normal, Plackett, and student’s-T copulas) and Gaussian kernel in conjunction with conventional as well as BGKE bandwidth estimators. Sensitivity analysis indicated that selection of optimum number of bins is critical in implementing D-kde in bivariate setting. In the context of univariate regional flood frequency analysis (RFFA) considering peak flows, a methodology based on D-kde and Index-flood methods is proposed and its performance is shown to be better when compared to that of widely used L-moment and Index-flood based method (‘regional L-moment algorithm’) through Monte-Carlo simulation experiments on homogeneous as well as heterogeneous synthetic regions, and through leave-one-out cross validation experiment performed on data sets pertaining to 54 watersheds in Godavari river basin, India. In this context, four homogeneous groups of watersheds are delineated in Godavari river basin using kernel principal component analysis (KPCA) in conjunction with Fuzzy c-means cluster analysis in L-moment framework, as an improvement over heterogeneous regions in the area (river basin) that are currently being considered by Central Water Commission, India. In the context of bivariate RFFA two methods are proposed. They involve forming site-specific pooling groups (regions) based on either L-moment based bivariate homogeneity test (R-BHT) or bivariate Kolmogorov-Smirnov test (R-BKS), and RFA based on D-kde. Their performance is assessed by application to data sets pertaining to stations in the conterminous United States. Results indicate that the R-BKS method is better than R-BHT in predicting quantiles of bivariate flood characteristics at ungauged sites, although the size of pooling groups formed using R-BKS is, in general, smaller than size of those formed using R-BHT. In general, the performance of the methods is found to improve with increase in size of pooling groups. Overall the results indicate that the D-kde always yields bona fide PDF (and CDF) in the context of univariate as well as bivariate flood frequency analysis, as probability density is nonnegative for all data points and integrates to unity for the valid range of the data. The performance of D-kde based at-site as well as regional FFA methodologies is found to be effective in univariate as well as bivariate settings, irrespective of the nature of population and sample size. A primary assumption underlying conventional FFA procedures has been that the time series of peak flow is stationarity (temporally homogeneous). However, recent studies carried out in various parts of the World question the assumption of flood stationarity. In this perspective, Time Varying Gaussian Copula (TVGC) based methodology is proposed in the thesis for flood frequency analysis in bivariate setting, which allows relaxing the assumption of stationarity in flood related variables. It is shown to be effective than seven commonly used stationary copulas through Monte-Carlo simulation experiments and by application to data sets pertaining to stations in the conterminous United States for which null hypothesis that peak flow data were non-stationary cannot be rejected.

Cette thèse est dédiée à l'analyse de stabilité des systèmes à pas d'échantillonnage variable et à la commande dynamique de l'échantillonnage. L'objectif est de concevoir des lois d'échantillonnage permettant de réduire la fréquence d'actualisation de la commande par retour d'état, tout en garantissant la stabilité du système.Tout d'abord, un aperçu des récents défis et axes de recherche sur les systèmes échantillonnés est présenté. Ensuite, une nouvelle approche de contrôle dynamique de l'échantillonnage, "échantillonnage dépendant de l'état", est proposée. Elle permet de concevoir hors-ligne un échantillonnage maximal dépendant de l'état défini sur des régions coniques de l'espace d'état, grâce à des LMIs.Plusieurs types de systèmes sont étudiés. Tout d'abord, le cas de système LTI idéal est considéré. La fonction d'échantillonnage est construite au moyen de polytopes convexes et de conditions de stabilité exponentielle de type Lyapunov-Razumikhin. Ensuite, la robustesse vis-à-vis des perturbations est incluse. Plusieurs applications sont proposées: analyse de stabilité robuste vis-à-vis des variations du pas d'échantillonnage, contrôles event-triggered et self-triggered, et échantillonnage dépendant de l'état. Enfin, le cas de système LTI perturbé à retard est traité. La construction de la fonction d'échantillonnage est basée sur des conditions de stabilité L2 et sur un nouveau type de fonctionnelles de Lyapunov-Krasovskii avec des matrices dépendant de l'état. Pour finir, le problème de stabilisation est traité, avec un nouveau contrôleur dont les gains commutent en fonction de l'état du système. Un co-design contrôleur/fonction d'échantillonnage est alors proposé / This PhD thesis is dedicated to the stability analysis of sampled-data systems with time-varying sampling, and to the dynamic control of the sampling instants. The main objective is to design sampling laws that allow for reducing the sampling frequency of state-feedback control for linear systems while ensuring the system's stability.First, an overview of the recent problems, challenges, and research directions regarding sampled-data systems is presented. Then, a novel dynamic sampling control approach, "state-dependent sampling", is proposed. It allows for designing offline a maximal state-dependent sampling map over conic regions of the state space, thanks to LMIs.Various classes of systems are considered throughout the thesis. First, we consider the case of ideal LTI systems, and propose a sampling map design based on the use of polytopic embeddings and Lyapunov-Razumikhin exponential stability conditions. Then, the robustness with respect to exogenous perturbations is included. Different applications are proposed: robust stability analysis with respect to time-varying sampling, as well as event-triggered, self-triggered, and state-dependent sampling control schemes. Finally, a sampling map design is proposed in the case of perturbed LTI systems with delay in the feedback control loop. It is based on L2-stability conditions and a novel type of Lyapunov-Krasovskii functionals with state-dependent matrices. Here, the stabilization issue is considered, and a new controller with gains that switch according to the system's state is presented. A co-design controller gains/sampling map is then proposed

La tesi di dottorato, composta da tre capitoli, si concentra sulla valutazione delle dinamiche di inflazione in Georgia e sulla previsione dei tassi di cambio nominali per i Paesi della European Eastern Partnership attraverso l’utilizzo di moderne tecniche econometriche. Nel primo capitolo, abbiamo svolto un’indagine sui modelli di previsione dei tassi di cambio e dell’inflazione. Questa indagine rivela che i modelli “factor-based and time-varying parameter” generano migliori previsioni rispetto ad altri modelli. Nel secondo capitolo, abbiamo approfondito le dinamiche di inflazione in Georgia utilizzando la New Keynesian Phillips Curve ibrida, inserita all’interno di un quadro di un modello “time-varying parameter (TVP)”. Una stima del modello TVP con volatilità stocastica mostra la persistenza di un’inflazione bassa durante il periodo 1996-2012. Un’analisi più approfondita dal 2003 mostra una volatilità crescente dell’inflazione. Inoltre, le stime del parametro evidenziano che la componente forward-looking del modello è importante a seguito dell’adozione di inflation targeting da parte della NBG a partire dal 2009. Nel terzo capitolo, abbiamo costruito dei modelli fattoriali, “Factor Vector Autoregressive” per prevedere i tassi di cambio nominali per i Paesi dell’European Eastern Partnership. Questi modelli prevedono meglio i tassi di cambio nominali rispetto ad un processo naïve come il random walk. / The Ph.D. thesis consist of three chapters on evaluating inflation dynamics in Georgia and modeling and forecasting nominal exchange rates for the European Eastern Partnership (EaP) countries using modern applied econometric techniques. In the first chapter, we survey of models those produce high predictive powers for forecasting exchange rates and inflation. This survey reveals that the factor-based and time-varying parameter (TVP) models generate superior forecasts relative to all other models. In the second chapter, we study inflation dynamics in Georgia using a hybrid New Keynesian Phillips Curve (NKPC) nested within a time-varying parameter (TVP) framework. Estimation of a TVP model with stochastic volatility shows low inflation persistence over the entire time span (1996-2012), while revealing increasing volatility of inflation shocks since 2003. Moreover, parameter estimates point to the forward-looking component of the model gaining importance following the National Bank of Georgia (NBG) adoption of inflation targeting in 2009. In the third chapter, we construct Factor Vector Autoregressive (FVAR) models to forecast nominal exchange rates for the EaP countries. This study provides better forecasts of nominal exchange rates than those produced by the random walk process.

Stavové modely jsou neobyčejně užitečné v mnoha inženýrských a vědeckých oblastech. Jejich atraktivita vychází především z toho faktu, že poskytují obecný nástroj pro popis široké škály dynamických systémů reálného světa. Nicméně, z důvodu jejich obecnosti, přidružené úlohy inference parametrů a stavů jsou ve většině praktických situacích nepoddajné. Tato dizertační práce uvažuje dvě zvláště důležité třídy nelineárních a ne-Gaussovských stavových modelů: podmíněně konjugované stavové modely a Markovsky přepínající nelineární modely. Hlavní rys těchto modelů spočívá v tom, že---navzdory jejich nepoddajnosti---obsahují poddajnou podstrukturu. Nepoddajná část požaduje abychom využily aproximační techniky. Monte Carlo výpočetní metody představují teoreticky a prakticky dobře etablovaný nástroj pro řešení tohoto problému. Výhoda těchto modelů spočívá v tom, že poddajná část může být využita pro zvýšení efektivity Monte Carlo metod tím, že se uchýlíme k Rao-Blackwellizaci. Konkrétně, tato doktorská práce navrhuje dva Rao-Blackwellizované částicové filtry pro identifikaci buďto statických anebo časově proměnných parametrů v podmíněně konjugovaných stavových modelech. Kromě toho, tato práce adoptuje nedávnou particle Markov chain Monte Carlo metodologii pro návrh Rao-Blackwellizovaných částicových Gibbsových jader pro vyhlazování stavů v Markovsky přepínajících nelineárních modelech. Tyto jádra jsou posléze použity pro inferenci parametrů metodou maximální věrohodnosti v uvažovaných modelech. Výsledné experimenty demonstrují, že navržené algoritmy překonávají příbuzné techniky ve smyslu přesnosti odhadu a výpočetního času.