We study energy-efficient power allocation among relays for lifetime maximization in a dual-hop relay network operated by amplify-and-forward relays with battery limitations. Power allocation algorithms are proposed for three different scenarios. First, we study the relay cooperation case where all the relays jointly support transmissions for a targeted data rate. By exploring the correlation of time-varying relay channels, we develop a prediction-based relay cooperation method for optimal power allocation strategy to improve the relay network lifetime over existing methods that do not predict the future channel state, or assume the current channel state remains static in the future. Next, we consider energy-efficient relay selection for the single source-destination case. Assuming finite transmission power levels, we propose a stochastic shortest path approach which gives the optimal relay selection decision to maximize the network lifetime. Due to the high computational complexity, a suboptimal prediction-based relay selection algorithm, directly coming from previous problem, is created. Finally, we extend our study to multiple source-destination case, where relay selection needs to be determined for each source-destination pair simultaneously. The network lifetime in the presence of multiple source-destination pairs is defined as the longest time when all source-destination pairs can maintain the target transmission rate. We design relay-to-destination mapping algorithms to prolong the network lifeii time. They all aim at maximizing the perceived network lifetime at the current time slot. The optimal max-min approach and suboptimal user-priority based approach are proposed with different levels of computational complexity. / UOIT

Le problème principalement étudié dans ce manuscrit est la stabilisation d'orbites périodiques de systèmes dynamiques non linéaires à l'aide d'une commande de rétroaction (feedback). Le but des méthodes de contrôle proposées ici est d'obtenir une oscillation périodique stable. Ces méthodes de contrôle sont appliquées à des systèmes présentant des orbites périodiques instables dans l'espace d'état, et ces dernières sont les orbites destinées à être stabilisées. Les méthodes proposées ici sont telles que l'oscillation stable qui en résulte est obtenue avec un effort de contrôle faible, et que la valeur de la commande tend vers zéro lorsque la trajectoire tend vers l'orbite stabilisée. La stabilité locale des orbites périodiques est analysée par l'étude de la stabilité des systèmes linéaires périodiques à l'aide de la théorie de Floquet. Ces systèmes linéaires sont obtenus par linéarisation des trajectoires au voisinage de l'orbite périodique. Les méthodes de contrôle utilisées ici pour la stabilisation des orbites périodiques sont une loi de commande proportionnelle, une loi de commande de rétroaction retardée et une loi de commande de rétroaction basée sur une prédiction. Ces méthodes sont appliquées aux systèmes en temps discret et aux systèmes en temps continu avec les modifications nécessaires. Les contributions principales de cette thèse sont associées à ces méthodes, proposant une méthode alternative de design de gain, une nouvelle loi de commande et des résultats associés.

Dynamic real-time distributed systems are characterized by significant run-time uncertainties at the mission and system levels. Typically, processing and communication latencies in such systems do not have known upper bounds and event and task arrivals and failure occurrences are non-deterministically distributed. This thesis proposes adaptive resource management heuristic techniques for periodic tasks in dynamic real-time distributed systems with the (soft real-time) objective of minimizing missed deadline ratios. The proposed resource management techniques continuously monitor the application tasks at run-time for adherence to the desired real-time requirements, detects timing failures or trends for impending failures (due to workload fluctuations), and dynamically allocate resources by replicating subtasks of application tasks for load sharing. We present "predictive" resource allocation algorithms that determine the number of subtask replicas that are required for adapting the application to a given workload situation using statistical regression theory. The algorithms use regression equations that forecast subtask timeliness as a function of external load parameters such as number of sensor reports and internal resource load parameters such as CPU utilization. The regression equations are determined off-line and on-line from application profiles that are collected off-line and on-line, respectively. To evaluate the performance of the predictive algorithms, we consider algorithms that determine the number of subtask replicas using empirically determined functions. The empirical functions compute the number of replicas as a function of the rate of change in the application workload during a "window" of past task periods. We implemented the resource management algorithms as part of a middleware infrastructure and measured the performance of the algorithms using a real-time benchmark. The experimental results indicate that the predictive, regression theory-based algorithms generally produce lower missed deadline ratios than the empirical strategies under the workload conditions that were studied. / Master of Science

Le problème principalement étudié dans ce manuscrit est la stabilisation d’orbites périodiques de systèmes dynamiques non linéaires à l’aide d’une commande de rétroaction (feedback). Le but des méthodes de contrôle proposées ici est d’obtenir une oscillation périodique stable. Ces méthodes de contrôle sont appliquées à des systèmes présentant des orbites périodiques instables dans l’espace d’état, et ces dernières sont les orbites destinées à être stabilisées.Les méthodes proposées ici sont telles que l’oscillation stable qui en résulte est obtenue avec un effort de contrôle faible, et que la valeur de la commande tend vers zéro lorsque la trajectoire tend vers l’orbite stabilisée. La stabilité locale des orbites périodiques est analysée par l’étude de la stabilité des systèmes linéaires périodiques à l’aide de la théorie de Floquet. Ces systèmes linéaires sont obtenus par linéarisation des trajectoires au voisinage de l’orbite périodique.Les méthodes de contrôle utilisées ici pour la stabilisation des orbites périodiques sont une loi de commande proportionnelle, une loi de commande de rétroaction retardée et une loi de commande de rétroaction basée sur une prédiction. Ces méthodes sont appliquées aux systèmes en temps discret et aux systèmes en temps continu avec les modifications nécessaires. Les contributions principales de cette thèse sont associées à ces méthodes, proposant une méthode alternative de design de gain, une nouvelle loi de commande et des résultats associés. / The main problem evaluated in this manuscript is the stabilization of periodic orbits of non-linear dynamical systems by use of feedback control. The goal of the control methods proposed in this work is to achieve a stable periodic oscillation. These control methods are applied to systems that present unstable periodic orbits in the state space, and the latter are the orbits to be stabilized.The methods proposed here are such that the resulting stable oscillation is obtained with low control effort, and the control signal is designed to converge to zero when the trajectory tends to the stabilized orbit. Local stability of the periodic orbits is analyzed by studying the stability of some linear time-periodic systems, using the Floquet stability theory. These linear systems are obtained by linearizing the trajectories in the vicinity of the periodic orbits.The control methods used for stabilization of periodic orbits here are the proportional feedback control, the delayed feedback control and the prediction-based feedback control. These methods are applied to discrete and continuous-time systems with the necessary modifications. The main contributions of the thesis are related to these methods, proposing an alternative control gain design, a new control law and related results. / O principal problema avaliado neste manuscrito é a estabilização de órbitas periódicas em sistemas dinâmicos não-lineares utilizando controle por realimentação. O objetivo dos métodos de controle propostos neste trabalho é obter uma oscilação periódica estável. Estes métodos de controle são aplicados a sistemas que apresentam órbitas periódicas instáveis no espaço de estados, estas são as órbitas a serem estabilizadas.Os métodos propostos aqui são tais que a oscilação periódica estável resultante é obtida utilizando um baixo esforço de controle, e o sinal de controle é projetado de forma a convergir para zero quanto a trajetória tende à órbita estabilizada. A estabilidade local de órbitas periódicas é analisada através do estudo da estabilidade de alguns sistemas lineares periódicos no tempo, utilizando a teoria de estabilidade de Floquet. Estes sistemas lineares são obtidos por linearização das trajetórias na vizinhança da órbita periódica.Os métodos de controle utilizados aqui para estabilização de órbitas periódicas são o proportional feedback control, o delayed feedback control e o prediction-based feedback control (controle por realimentação baseado em predição). Estes métodos são aplicados a sistemas de tempo discreto e de tempo contínuo, com as modificações necessárias. As principais contribuições da tese são relacionadas a esses métodos, propondo um projeto de ganho de controle alternativo, uma nova lei de controle e resultados relacionados.

The main problem evaluated in this manuscript is the stabilization of periodic orbits of non-linear dynamical systems by use of feedback control. The goal of the control methods proposed in this work is to achieve a stable periodic oscillation. These control methods are applied to systems that present unstable periodic orbits in the state space, and the latter are the orbits to be stabilized.The methods proposed here are such that the resulting stable oscillation is obtained with low control effort, and the control signal is designed to converge to zero when the trajectory tends to the stabilized orbit. Local stability of the periodic orbits is analyzed by studying the stability of some linear time-periodic systems, using the Floquet stability theory. These linear systems are obtained by linearizing the trajectories in the vicinity of the periodic orbits.The control methods used for stabilization of periodic orbits here are the proportional feedback control, the delayed feedback control and the prediction-based feedback control. These methods are applied to discrete and continuous-time systems with the necessary modifications. The main contributions of the thesis are related to these methods, proposing an alternative control gain design, a new control law and related results.

Diplomová práce pojednává o stabilizaci diskrétního logistického modelu pomocí několika řídících metod. Je zde provedena především stabilizace rovnováh, 2-periodických cyklů a 3-periodických cyklů. Ke stabilizaci systému je využito proporčního zpětně-vazebního řízení, zpětně-vazebního řízení s časovým zpožděním a řízení založeného na predikci. U každé metody je diskutovaná stabilizační množina pro řídící zesilovač spolu s oblastmi stability pro odpovídající kontrolovaná řešení. Všechny teoretické výsledky jsou ilustrovány grafickými interpretacemi v softwaru MATLAB. Podpůrné výpočty jsou provedeny pomocí softwaru Maple.