The extensive use of a least-squares problem formulation in many fields is partly motivated by the existence of an analytic solution formula which makes the theory comprehensible and readily applicable, but also easily embedded in computer-aided design or analysis tools. While the mathematics behind convex optimization has been studied for about a century, several recent researches have stimulated a new interest in the topic. Convex optimization, being a special class of mathematical optimization problems, can be considered as generalization of both least-squares and linear programming. As in the case of a linear programming problem there is in general no simple analytical formula that can be used to find the solution of a convex optimization problem. There exists however efficient methods or software implementations for solving a large class of convex problems. The challenge and the state of the art in using convex optimization comes from the difficulty in recognizing and formulating the problem. The main goal of this thesis is to investigate the potential advantages and benefits of convex optimization techniques in the field of system identification. The primary work focuses on parametric discrete-time system identification models in which we assume or choose a specific model structure and try to estimate the model parameters for best fit using experimental input-output (IO) data. By developing a working knowledge of convex optimization and treating the system identification problem as a convex optimization problem will allow us to reduce the uncertainties in the parameter estimation. This is achieved by reecting prior knowledge about the system in terms of constraint functions in the least-squares formulation.

As we begin to analyze and construct extremely complex systems, a theory for understanding and designing the underlying architecture becomes very important. To move in the direction of a precise theory of architecture, this thesis will provide some concrete tools to analyze and design complex systems with a given network structure. The first main result of this thesis analyzes the vulnerability of a system and shows that a system's vulnerability depends on its network structure. We will consider destabilization attacks acting on a single link in a system's logical interconnection structure. The concept of a vulnerable link is characterized and necessary and sufficient conditions for identifying these links are provided. The vulnerability of various system architectures are then characterized by the vulnerability of their weakest link, and it is shown that every transfer function has a completely secure architecture with no vulnerable links. The second part of this thesis focuses on synthesizing controllers with a specified network structure. It presents a new approach to distributed controller design that exploits the dynamical structure function representation of linear time invariant systems to characterize the structure of a system. The design technique sequentially constructs each link in an arbitrary controller signal structure, and the main theorem proves that either the resulting controller is stabilizing or that no controller with the desired structure can stabilize the system.

A backbone network is the central part of the communication network, which provides connectivity within the various systems across large distances. Disruptions in a backbone network would cause severe consequences which could manifest in the service outage on a large scale. Depending on the size and the importance of the network, its failure could leave a substantial impact on the area it is associated with. The failures of the network services could lead to a significant disturbance of human activities. Therefore, making backbone communication networks more resilient directly affects the resilience of the area. Contemporary urban and regional development overwhelmingly converges with the communication infrastructure expansion and their obvious mutual interconnections become more reciprocal. Spreading failures are of particular interest. They usually originate in a single network segment and then spread to the rest of network often causing a global collapse. Two types of spreading failures are given focus, namely: epidemics and cascading failures. How to make backbone networks more resilient against spreading failures? How to tune the topology or additionally protect nodes or links in order to mitigate an effect of the potential failure? Those are the main questions addressed in this thesis. First, the epidemic phenomena are discussed. The subjects of epidemic modeling and identification of the most influential spreaders are addressed using a proposed Linear Time-Invariant (LTI) system approach. Throughout the years, LTI system theory has been used mostly to describe electrical circuits and networks. LTI is suitable to characterize the behavior of the system consisting of numerous interconnected components. The results presented in this thesis show that the same mathematical toolbox could be used for the complex network analysis. Then, cascading failures are discussed. Like any system which can be modeled using an interdependence graph with limited capacity of either nodes or edges, backbone networks are prone to cascades. Numerical simulations are used to model such failures. The resilience of European National Research and Education Networks (NREN) is assessed, weak points and critical areas of the network are identified and the suggestions for its modification are proposed.

A backbone network is the central part of the communication network, which provides connectivity within the various systems across large distances. Disruptions in a backbone network would cause severe consequences which could manifest in the service outage on a large scale. Depending on the size and the importance of the network, its failure could leave a substantial impact on the area it is associated with. The failures of the network services could lead to a significant disturbance of human activities. Therefore, making backbone communication networks more resilient directly affects the resilience of the area. Contemporary urban and regional development overwhelmingly converges with the communication infrastructure expansion and their obvious mutual interconnections become more reciprocal. Spreading failures are of particular interest. They usually originate in a single network segment and then spread to the rest of network often causing a global collapse. Two types of spreading failures are given focus, namely: epidemics and cascading failures. How to make backbone networks more resilient against spreading failures? How to tune the topology or additionally protect nodes or links in order to mitigate an effect of the potential failure? Those are the main questions addressed in this thesis. First, the epidemic phenomena are discussed. The subjects of epidemic modeling and identification of the most influential spreaders are addressed using a proposed Linear Time-Invariant (LTI) system approach. Throughout the years, LTI system theory has been used mostly to describe electrical circuits and networks. LTI is suitable to characterize the behavior of the system consisting of numerous interconnected components. The results presented in this thesis show that the same mathematical toolbox could be used for the complex network analysis. Then, cascading failures are discussed. Like any system which can be modeled using an interdependence graph with limited capacity of either nodes or edges, backbone networks are prone to cascades. Numerical simulations are used to model such failures. The resilience of European National Research and Education Networks (NREN) is assessed, weak points and critical areas of the network are identified and the suggestions for its modification are proposed.

La tendance actuelle dans le domaine des télécommunications mène à des systèmes capables de fonctionner selon plusieurs standards, et donc plusieurs fréquences porteuses. La synthèse de la fréquence porteuse est un élément clef, dont les propriétés reposent essentiellement sur les performances de l’oscillateur employé. Pour assurer le fonctionnement de systèmes compatibles avec plusieurs standards de télécommunication, la solution conventionnelle consiste à intégrer plusieurs oscillateurs locaux. Cette solution est coûteuse, d’autant plus que, malgré le fait que les technologies actuelles atteignent des niveaux d’intégration très importants, la surface occupée par des oscillateurs traditionnels de type LC ne peut pas être diminuée, alors que le coût de fabrication au millimètre carré devient de plus en plus élevé. Il serait donc très intéressant de remplacer les oscillateurs LC, ce qui nous amène à rechercher des solutions alternatives parmi de nouvelles technologies. L’oscillateur spintronique (STO) est un nouveau dispositif issu des études sur les couches minces magnétiques. Il apparait comme un candidat potentiel de remplacement des oscillateurs LC du fait de sa grande accordabilité en fréquence et de son faible encombrement. Toutefois des mesures effectuées sur les STOs ont montré que la performance en puissance et en bruit de phase d’un oscillateur seul ne permet pas de remplir les spécifications pour des applications de télécommunication. Nous proposons de remplir ces spécifications en couplant un nombre d’oscillateurs spintroniques important. Dans ce cadre se posent plusieurs questions qui concernent les procédures de modélisation, d’analyse et de synthèse des systèmes interconnectés. Les procédures de modélisation incluent la démarche de recherche de modèles à complexité croissante qui décrivent les propriétés entrée-sortie d’un oscillateur spintronique, ainsi que la démarche de généralisation des modèles des oscillateurs dans le cadre du réseau. Les procédures d’analyse cherchent à vérifier la stabilité et évaluer la performance des systèmes interconnectés. Les procédures de synthèse permettent de concevoir des interconnexions sophistiquées pour les oscillateurs afin d’assurer toutes les spécifications du cahier des charges. Dans ce document, nous établissons tout d’abord le problème de la synthèse de fréquence par couplage avec un cahier des charges formalisé en termes de gabarits fréquentiels sur des densités spectrales de puissance. Le cahier des charges posé amène la nécessité de modéliser l’oscillateur spintronique pour pouvoir simuler et analyser son comportement. Ici, nous proposons une modélisation originale selon des degrés de complexité croissante. Ensuite, nous discutons de la structure de la commande de l’ensemble des oscillateurs afin de remplir les spécifications du cahier des charges. La structure de commande proposée nécessite de développer une méthode de conception des interconnexions du réseau d’après les critères de performance. Dans les deux derniers chapitres, nous proposons deux méthodes fréquentielles de synthèse originales pour résoudre le problème de synthèse de fréquence par couplage. La première méthode de synthèse permet de prendre en compte un critère mathématique du cahier des charges, qui correspond à un gabarit fréquentiel à respecter, et permet d’obtenir une matrice d’interconnexion des sous-systèmes, telle que le module de la réponse fréquentielle du réseau approxime le gabarit imposé par le cahier des charges. La deuxième méthode de synthèse permet de prendre en compte plusieurs gabarits fréquentiels à la fois. La solution obtenue est une matrice d’interconnexion des sous-systèmes, qui résout le problème de la synthèse de fréquence par couplage d’oscillateurs spintroniques. / The current trends in telecommunication are leading to systems that are compatible with multiple standards and consequently multiple carrier frequencies. The frequency synthesis is a key element influenced by the local oscillator performance. In order to ensure the system compatibility with multiple telecommunication standards, the conventional solution consists in using one local oscillator for each standard. This solution is expensive, even more, since the cost per squared millimetre is increasing, while the silicon area occupied by the traditional LC-tank oscillators cannot be reduced in spite of the fact that technology is going to higher integration levels. Thus, it should be interesting to find a substitution to the LC-tank oscillators which leads to research for alternative solutions among new technologies. The spin torque oscillator (STO) is a new device issued from the ferromagnetic thin-film research. Due to its frequency accord ability and its capability to occupy relatively small volume, it appears as a potential candidate for the LC-tank oscillator replacement. However, a set of measurements prove that these devices exhibit poor power and phase noise performance, making them unable to fulfill the technical specification of the radiofrequency applications. We propose to reach these specifications by coupling of a large number of spin torque oscillators. In this scope, numerous questions appear regarding the procedures of modelling, analysis and synthesis of the complex interconnected systems. The modelling procedures are dedicated to the increasing complexity models that describe the input-output behaviour of a spin torque oscillator and its behaviour within the interconnected network. The analysis procedures are targeted to verify the stability and to evaluate the performance level of the interconnected systems. The synthesis procedures allow to design the interconnection law for spin torque oscillators in order to fulfill the technical requirements. In this document, the frequency synthesis problem by spin torque oscillator coupling with technical specification description in terms of power spectral densities is established. The formulated specifications introduce the problem of the oscillator modelling in order to perform a simulation and an analysis of the oscillator behaviour. Here, we propose an original model using several conventional models with increasing complexity. An original oscillator network model that describes qualitative properties of the oscillator synchronisation is introduced. Afterwards, the control law architecture for an oscillator set is established in order to accomplish the technical requirement specifications. The suggested control architecture needs to be developed with quantitative systematic and efficient design method for the network interconnection taking into account the formulated performance criteria. In the last two chapters we propose two original frequency domain design methods allowing the resolution of our frequency synthesis problem. The first design method allows to consider explicitly a performance criterium corresponding toa desired frequency constraint. The method allows to obtain a suitable sub-system interconnection matrix that fits the frequency specification constraint. The second design method allows to find an interconnection matrix and to take into account simultaneously several frequency specification constraints. The interconnection matrix obtained with the proposed method solves the problem of frequency synthesis by coupling of spin torque oscillators.