abstract: This research mainly focuses on improving the utilization of photovoltaic (PV) re-sources in distribution systems by reducing their variability and uncertainty through the integration of distributed energy storage (DES) devices, like batteries, and smart PV in-verters. The adopted theoretical tools include statistical analysis and convex optimization. Operational issues have been widely reported in distribution systems as the penetration of PV resources has increased. Decision-making processes for determining the optimal allo-cation and scheduling of DES, and the optimal placement of smart PV inverters are con-sidered. The alternating current (AC) power flow constraints are used in these optimiza-tion models. The first two optimization problems are formulated as quadratically-constrained quadratic programming (QCQP) problems while the third problem is formu-lated as a mixed-integer QCQP (MIQCQP) problem. In order to obtain a globally opti-mum solution to these non-convex optimization problems, convex relaxation techniques are introduced. Considering that the costs of the DES are still very high, a procedure for DES sizing based on OpenDSS is proposed in this research to avoid over-sizing. Some existing convex relaxations, e.g. the second order cone programming (SOCP) relaxation and semidefinite programming (SDP) relaxation, which have been well studied for the optimal power flow (OPF) problem work unsatisfactorily for the DES and smart inverter optimization problems. Several convex constraints that can approximate the rank-1 constraint X = xxT are introduced to construct a tighter SDP relaxation which is referred to as the enhanced SDP (ESDP) relaxation using a non-iterative computing framework. Obtaining the convex hull of the AC power flow equations is beneficial for mitigating the non-convexity of the decision-making processes in power systems, since the AC power flow constraints exist in many of these problems. The quasi-convex hull of the quadratic equalities in the AC power bus injection model (BIM) and the exact convex hull of the quadratic equality in the AC power branch flow model (BFM) are proposed respectively in this thesis. Based on the convex hull of BFM, a novel convex relaxation of the DES optimizations is proposed. The proposed approaches are tested on a real world feeder in Arizona and several benchmark IEEE radial feeders. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2016

abstract: The problem of systematically designing a control system continues to remain a subject of intense research. In this thesis, a very powerful control system design environment for Linear Time-Invariant (LTI) Multiple-Input Multiple-Output (MIMO) plants is presented. The environment has been designed to address a broad set of closed loop metrics and constraints; e.g. weighted H-infinity closed loop performance subject to closed loop frequency and/or time domain constraints (e.g. peak frequency response, peak overshoot, peak controls, etc.). The general problem considered - a generalized weighted mixed-sensitivity problem subject to constraints - permits designers to directly address and tradeoff multivariable properties at distinct loop breaking points; e.g. at plant outputs and at plant inputs. As such, the environment is particularly powerful for (poorly conditioned) multivariable plants. The Youla parameterization is used to parameterize the set of all stabilizing LTI proper controllers. This is used to convexify the general problem being addressed. Several bases are used to turn the resulting infinite-dimensional problem into a finite-dimensional problem for which there exist many efficient convex optimization algorithms. A simple cutting plane algorithm is used within the environment. Academic and physical examples are presented to illustrate the utility of the environment. / Dissertation/Thesis / M.S. Electrical Engineering 2013

Apresentamos dois novos métodos para a solução de problemas de otimização convexa em dois níveis não necessariamente diferenciáveis, i.e., mostramos que as sequências geradas por ambos os métodos convergem para o conjunto ótimo de uma função não suave sujeito a um conjunto que também envolve a minimização de uma função não diferenciável. Ambos os algoritmos dispensam qualquer tipo de resolução de subproblemas ou busca linear durante suas iterações. Ao final, para demonstrar que os métodos são viáveis, resolvemos um problema de reconstrução de imagens tomográficas / We present two new methods for solving bilevel convex optimization problems, where both functions are not necessarily differentiable, i.e., we show that the sequences generated by those methods converge to the optimal set of a nonsmooth function subject to a set that also involves a function minimization. Both algorithms do not require any kind of subproblems resolution or linear search during the iterations. At the end, to prove that our methods are viable, we solve a problem of tomographic image reconstruction

Neste trabalho apresentamos um novo método ótimo para otimização de uma função convexa diferenciável sujeita a restrições convexas. Nosso método é baseado em ideias de Nesterov e Auslender e Teboulle. A proposta dos últimos autores usa uma distância de Bregman coerciva para garantir que os iterados permaneçam no interior do conjunto viável. Nosso método estende esses resultados para permitir o emprego da distância Euclidiana ao quadrado. Mostramos também como estimar a constante de Lipschitz para o gradiente da função objetivo, o que resulta em uma melhora na eficiência numérica do método. Finalmente, apresentamos experimentos numéricos para validar nossa proposta e comparar com o algoritmo de Nesterov. / In this work we introduce a new optimal method for constrained differentiable convex optimization which is based on previous ideas by Nesterov and Auslender and Teboulle. The method proposed by the last authors use a coercive Bregman distance to ensure that the iterates remain in the interior of the feasible set. Our results extend this method to allow the use of the squared Euclidean distance. We also show how to estimate the Lipschitz constant of the gradient of the objective function, improving the numerical behavior of the method. Finally, we present numerical experiments to validate our approach and compare it to Nesterov\'s algorithm.

Este trabalho apresenta uma arquitetura de controle de movimento entre duas posturas distintas para um robô móvel sob rodas com acionamento diferencial em um ambiente estruturado e livre de obstáculos. O conceito clássico de eficiência foi utilizado para a definição das estratégias de controle: um robô se movimenta de forma eficiente quando realiza a tarefa determinada no menor tempo e utilizando menor quantidade energética. A arquitetura proposta é um recorte do modelo de Controle Hierárquico Aninhado (NHC), composto por três níveis de abstração: (i) Planejamento de Caminho, (ii) Planejamento de Trajetória e (iii) Rastreamento de Trajetória. O Planejamento de Caminho proposto suaviza uma geodésica Dubins - o caminho mais eficiente - por uma Spline Grampeada para que este caminho seja definido por uma curva duplamente diferenciável. Uma transformação do espaço de configuração do robô é realizada. O Planejamento de Trajetória é um problema de otimização convexa na forma de Programação Cônica de Segunda Ordem, cujo objetivo é uma função ponderada entre tempo e energia. Como o tempo de percurso e a energia total consumida pelo robô possui uma relação hiperbólica, um algoritmo de sintonia do coeficiente de ponderação entre estas grandezas é proposta. Por fim, um Rastreador de Trajetória de dupla malha baseado em linearização entrada-saída e controle PID é proposto, e obteve resultados satisfatórios no rastreamento do caminho pelo robô. / This work presents a motion control architecture between two different positions for a differential driven wheeled mobile robot in a obstacles free structured environment. The classic concept of efficiency was used to define the control strategies: a robot moves efficiently when it accomplishes the determined task in the shortest time and using less amount of energy. The proposed architecture is a clipping of the Nested Hierarchical Controller (NHC) model, composed of three levels of abstraction: (i) Path Planning, (ii) Trajectory Planning and (iii) Trajectory Tracking. The proposed Path Planning smoothes a geodesic Dubins - the most efficient path - by a Clamped Spline as this path is defined by a twice differentiable curve. A transformation of the robot configuration space is performed. The Trajectory Planning is a convex optimization problem in the form of Second Order Cone Programming, whose objective is a weighted function between time and energy. As the travel time and the total energy consumed by the robot has a hyperbolic relation, a tuning algorithm to the weighting is proposed. Finnaly, a dual-loop Trajectory Tracker based on input-output feedback linearization and PID control is proposed, which obtained satisfactory results in tracking the path by the robot.

Un système est dit robuste s'il est possible de garantir son bon comportement dynamique malgré les dispersions de ses caractéristiques lors de sa fabrication, les variations de l'environnement ou encore son vieillissement. Au-delà du fait que la dispersion des caractéristiques est inéluctable, une plus grande dispersion permet notamment de diminuer fortement les coûts de production. La prise en compte explicite de la robustesse par les ingénieurs est donc un enjeu crucial lors de la conception d'un système. Des propriétés robustes peuvent être garanties lors de la synthèse d'un correcteur en boucle fermée. Il est en revanche beaucoup plus difficile de garantir ces propriétés en boucle ouverte, ce qui concerne par exemple des cas comme la synthèse d'estimateur.Prendre en compte la robustesse lors de la synthèse est une problématique importante de la communauté du contrôle robuste. Un certain nombre d'outils ont été développés pour analyser la robustesse d'un système vis-à-vis d'un ensemble d'incertitudes(μ analyse par exemple). Bien que le problème soit intrinsèquement complexe au sens algorithmique, des relaxations ont permis de formuler des conditions suffisantes pour tester la stabilité d'un système vis-à-vis d'un ensemble d'incertitudes. L'émergence de l'Optimisation sous contrainte Inégalité Matricielle Linéaire (LMI) a permis de tester ces conditions suffisantes au moyen d'un algorithme efficace, c'est-à-dire convergeant vers une solution en un temps raisonnable grâce au développement des méthodes des points intérieurs.En se basant sur ces résultats d'analyse, le problème de synthèse de correcteurs en boucle fermée ne peut pas être formulé sous la forme d'un problème d'optimisation pour lequel un algorithme efficace existe. En revanche, pour certains cas comme la synthèse de filtres robustes, le problème de synthèse peut être formulé sous la forme d'un problème d'optimisation sous contrainte LMI pour lequel un algorithme efficace existe. Ceci laisse entrevoir un certain potentiel de l'approche robuste pour la synthèse d'estimateurs.Exploitant ce fait, cette thèse propose une approche complète du problème de synthèse d'estimateurs robustes par l'intermédiaire des outils d'analyse de la commande robuste en conservant le caractère efficace de la synthèse lié aux outils classiques. Cette approche passe par une ré-interprétation de l'estimation nominale (sans incertitude) par l'optimisation sous contrainte LMI, puis par une extension systématique des outils de synthèse et d'analyse développés pour l'estimation nominale à l'estimation robuste.Cette thèse présente des outils de synthèse d'estimateurs, mais également des outils d'analyse qui permettront de tester les performances robustes atteintes par les estimateurs.Les résultats présentés dans ce document sont exprimés sous la forme de théorèmes présentant des contraintes LMI. Ces théorèmes peuvent se mettre de façon systématique sous la forme d'un problème d'optimisation pour lequel un algorithme efficace existe.Pour finir, les problèmes de synthèse d'estimateurs robustes appartiennent à une classe plus générale de problèmes de synthèse robuste : les problèmes de synthèse robuste en boucle ouverte. Ces problèmes de synthèse ont un potentiel très intéressant. Des résultats de base sont formulés pour la synthèse en boucle ouverte, permettant de proposer des méthodes de synthèse robustes dans des cas pour lesquels la mise en place d'une boucle de rétroaction est impossible. Une extension aux systèmes LPV avec une application à la commande de position sans capteur de position est également proposée. / A system is said to be robust if it is possible to guarantee his dynamic behaviour despite dispersion of his features due to production, environmental changes or aging. beyond the fact that a dispersion is ineluctable, a greater one allows to reduce production costs. Thus, considering robustness is a crucial stake during the conception of a system.Robustness can be achieved using feedback, but is more difficult in Open-Loop, which concerns estimator synthesis for instance.Robustness is a major concern of the Robust Control Community. Many tools have been developed to analyse robustness of a system towards a set of uncertainties (μ analysis for instance). And even if the problem is known to be difficult (speaking of complexity), sufficient conditions allow to formulate results to test the robust stability of a system. Thanks to the development of interior point methods, the emergence of optimization under Linear Matrix Inequalities Constraints allows to test these results using an efficient algorithm.Based on these analysis results, the robust controller synthesis problem cannot be recast as a convex optimization problem involving LMI. But for some cases such as filter synthesis, the synthesis problem can recast as a convex optimization problem. This fact let sense that robust control tools have some potential for estimators synthesis.Exploiting this fact, this thesis ofiers a complete approach of robust estimator synthesis, using robust control tools, while keeping what made the nominal approaches successful : eficient computation tools. this approach goes through reinterpretation of nominal estimation using LMI optimization, then propose a systematic extension of these tools to robust estimation.This thesis presents not only synthesis tools, but also analysis tools, allowing to test the robust performance reached by estimators All the results are proposed as convex optimization problems involving LMI.As a conclusion, robust estimator synthesis problems belong to a wider class of problems : robust open-loop synthesis problems, which have a great potential in many applications. Basic results are formulated for open-loop synthesis, providing results for cases where feedback cannot be used. An extension to LPV systems with an application to sensorless control is given.

Abstract Advanced multiple-input multiple-output (MIMO) transceiver structures which utilize the knowledge of channel state information (CSI) at the transmitter side to optimize certain link parameters (e.g., throughput, fairness, spectral efficiency, etc.) under different constraints (e.g., maximum transmitted power, minimum quality of services (QoS), etc.) are considered in this thesis. Adaptive transmission schemes for point-to-point MIMO systems are considered first. A robust link adaptation method for time-division duplex systems employing MIMO-OFDM channel eigenmode based transmission is developed. A low complexity bit and power loading algorithm which requires low signaling overhead is proposed. Two algorithms for computing the sum-capacity of MIMO downlink channels with full CSI knowledge are derived. The first one is based on the iterative waterfilling method. The convergence of the algorithm is proved analytically and the computer simulations show that the algorithm converges faster than the earlier variants of sum power constrained iterative waterfilling algorithms. The second algorithm is based on the dual decomposition method. By tracking the instantaneous error in the inner loop, a faster version is developed. The problem of linear transceiver design in MIMO downlink channels is considered for a case when the full CSI of scheduled users only is available at the transmitter. General methods for joint power control and linear transmit and receive beamformers design are provided. The proposed algorithms can handle multiple antennas at the base station and at the mobile terminals with an arbitrary number of data streams per scheduled user. The optimization criteria are fairly general and include sum power minimization under the minimum signal-to-interference-plus-noise ratio (SINR) constraint per data stream, the balancing of SINR values among data streams, minimum SINR maximization, weighted sum-rate maximization, and weighted sum mean square error minimization. Besides the traditional sum power constraint on the transmit beamformers, multiple sum power constraints can be imposed on arbitrary subsets of the transmit antennas.This extends the applicability of the results to novel system architectures, such as cooperative base station transmission using distributed MIMO antennas. By imposing per antenna power constraints, issues related to the linearity of the power amplifiers can be handled as well. The original linear transceiver design problems are decomposed as a series of remarkably simpler optimization problems which can be efficiently solved by using standard convex optimization techniques. The advantage of this approach is that it can be easily extended to accommodate various supplementary constraints such as upper and/or lower bounds for the SINR values and guaranteed QoS for different subsets of users. The ability to handle transceiver optimization problems where a network-centric objective (e.g., aggregate throughput or transmitted power) is optimized subject to user-centric constraints (e.g., minimum QoS requirements) is an important feature which must be supported by future broadband communication systems.

De nombreux domaines issus de l’industrie et des sciences appliquées ont été les témoins d’une révolution numérique. Cette dernière s’est accompagnée d’une croissance du volume des données, dont le traitement est devenu un défi technique. Dans ce contexte, la parcimonie est apparue comme un concept central en apprentissage statistique. Il est en effet naturel de vouloir exploiter les données disponibles via un nombre réduit de paramètres. Cette thèse se concentre sur une forme particulière et plus récente de parcimonie, nommée parcimonie structurée. Comme son nom l’indique, nous considérerons des situations où, au delà de la seule parcimonie, nous aurons également à disposition des connaissances a priori relatives à des propriétés structurelles du problème. L’objectif de cette thèse est d'analyser le concept de parcimonie structurée, en se basant sur des considérations statistiques, algorithmiques et appliquées. Nous commencerons par introduire une famille de normes structurées parcimonieuses dont les aspects statistiques sont étudiées en détail. Nous considérerons ensuite l’apprentissage de dictionnaires, où nous exploiterons les normes introduites précédemment dans un cadre de factorisation de matrices. Différents outils algorithmiques efficaces, tels que des méthodes proximales, seront alors proposés. Grâce à ces outils, nous illustrerons sur de nombreuses applications pourquoi la parcimonie structurée peut être bénéfique. Ces exemples contiennent des tâches de restauration en traitement de l’image, la modélisation hiérarchique de documents textuels, ou encore la prédiction de la taille d’objets à partir de signaux d’imagerie par résonance magnétique fonctionnelle. / Numerous fields of applied sciences and industries have been recently witnessing a process of digitisation. This trend has come with an increase in the amount digital data whose processing becomes a challenging task. In this context, parsimony, also known as sparsity, has emerged as a key concept in machine learning and signal processing. It is indeed appealing to exploit data only via a reduced number of parameters. This thesis focuses on a particular and more recent form of sparsity, referred to as structured sparsity. As its name indicates, we shall consider situations where we are not only interested in sparsity, but where some structural prior knowledge is also available. The goal of this thesis is to analyze the concept of structured sparsity, based on statistical, algorithmic and applied considerations. To begin with, we introduce a family of structured sparsity-inducing norms whose statistical aspects are closely studied. In particular, we show what type of prior knowledge they correspond to. We then turn to sparse structured dictionary learning, where we use the previous norms within the framework of matrix factorization. From an optimization viewpoint, we derive several efficient and scalable algorithmic tools, such as working-set strategies and proximal-gradient techniques. With these methods in place, we illustrate on numerous real-world applications from various fields, when and why structured sparsity is useful. This includes, for instance, restoration tasks in image processing, the modelling of text documents as hierarchy of topics, the inter-subject prediction of sizes of objects from fMRI signals, and background-subtraction problems in computer vision.

Les tâches qui consistent à comprendre automatiquement le contenu d’un signal naturel, comme une image ou un son, sont en général difficiles. En effet, dans leur représentation naïve, les signaux sont des objets compliqués, appartenant à des espaces de grande dimension. Représentés différemment, ils peuvent en revanche être plus faciles à interpréter. Cette thèse s’intéresse à une représentation fréquemment utilisée dans ce genre de situations, notamment pour analyser des signaux audio : le module de la transformée en ondelettes. Pour mieux comprendre son comportement, nous considérons, d’un point de vue théorique et algorithmique, le problème inverse correspondant : la reconstruction d’un signal à partir du module de sa transformée en ondelettes. Ce problème appartient à une classe plus générale de problèmes inverses : les problèmes de reconstruction de phase. Dans un premier chapitre, nous décrivons un nouvel algorithme, PhaseCut, qui résout numériquement un problème de reconstruction de phase générique. Comme l’algorithme similaire PhaseLift, PhaseCut utilise une relaxation convexe, qui se trouve en l’occurence être de la même forme que les relaxations du problème abondamment étudié MaxCut. Nous comparons les performances de PhaseCut et PhaseLift, en termes de précision et de rapidité. Dans les deux chapitres suivants, nous étudions le cas particulier de la reconstruction de phase pour la transformée en ondelettes. Nous montrons que toute fonction sans fréquence négative est uniquement déterminée (à une phase globale près) par le module de sa transformée en ondelettes, mais que la reconstruction à partir du module n’est pas stable au bruit, pour une définition forte de la stabilité. On démontre en revanche une propriété de stabilité locale. Nous présentons également un nouvel algorithme de reconstruction de phase, non-convexe, qui est spécifique à la transformée en ondelettes, et étudions numériquement ses performances. Enfin, dans les deux derniers chapitres, nous étudions une représentation plus sophistiquée, construite à partir du module de transformée en ondelettes : la transformée de scattering. Notre but est de comprendre quelles propriétés d’un signal sont caractérisées par sa transformée de scattering. On commence par démontrer un théorème majorant l’énergie des coefficients de scattering d’un signal, à un ordre donné, en fonction de l’énergie du signal initial, convolé par un filtre passe-haut qui dépend de l’ordre. On étudie ensuite une généralisation de la transformée de scattering, qui s’applique à des processus stationnaires. On montre qu’en dimension finie, cette transformée généralisée préserve la norme. En dimension un, on montre également que les coefficients de scattering généralisés d’un processus caractérisent la queue de distribution du processus. / Automatically understanding the content of a natural signal, like a sound or an image, is in general a difficult task. In their naive representation, signals are indeed complicated objects, belonging to high-dimensional spaces. With a different representation, they can however be easier to interpret. This thesis considers a representation commonly used in these cases, in particular for theanalysis of audio signals: the modulus of the wavelet transform. To better understand the behaviour of this operator, we study, from a theoretical as well as algorithmic point of view, the corresponding inverse problem: the reconstruction of a signal from the modulus of its wavelet transform. This problem belongs to a wider class of inverse problems: phase retrieval problems. In a first chapter, we describe a new algorithm, PhaseCut, which numerically solves a generic phase retrieval problem. Like the similar algorithm PhaseLift, PhaseCut relies on a convex relaxation of the phase retrieval problem, which happens to be of the same form as relaxations of the widely studied problem MaxCut. We compare the performances of PhaseCut and PhaseLift, in terms of precision and complexity. In the next two chapters, we study the specific case of phase retrieval for the wavelet transform. We show that any function with no negative frequencies is uniquely determined (up to a global phase) by the modulus of its wavelet transform, but that the reconstruction from the modulus is not stable to noise, for a strong notion of stability. However, we prove a local stability property. We also present a new non-convex phase retrieval algorithm, which is specific to the case of the wavelet transform, and we numerically study its performances. Finally, in the last two chapters, we study a more sophisticated representation, built from the modulus of the wavelet transform: the scattering transform. Our goal is to understand which properties of a signal are characterized by its scattering transform. We first prove that the energy of scattering coefficients of a signal, at a given order, is upper bounded by the energy of the signal itself, convolved with a high-pass filter that depends on the order. We then study a generalization of the scattering transform, for stationary processes. We show that, in finite dimension, this generalized transform preserves the norm. In dimension one, we also show that the generalized scattering coefficients of a process characterize the tail of its distribution.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第22862号 / 情博第741号 / 新制||情||127(附属図書館) / 京都大学大学院情報学研究科数理工学専攻 / (主査)教授 山下 信雄, 教授 太田 快人, 教授 鹿島 久嗣 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM