Montamos um aparato experimental para o estudo de comportamentos complexos na dinâmica de formação de gotas d\'água no bico de uma torneira. Desenvolvemos um sistema hidráulico em circuito fechado, e um sistema de aquisição de dados automatizado, que também controla a abertura da torneira (uma válvula de agulha). Utilizamos como parâmetro de controle a taxa de gotejamento estabelecida pela abertura da torneira. Os dados são séries de tempos {T n} entre gotas sucessivas para cada taxa de gotejamento. Utilizando diagramas de bifurcação, e reconstruções do espaço de fase com mapas de primeiro retomo Tn+1 x Tn , observamos duplicações de período, bifurcação de Hopf, crises interiores e de fronteira, comportamentos intermitentes, e movimentos quase-periódicos. Aplicamos anticontrole de caos, desestabilizando um ponto fixo estável com pulsos de ar comprimido sobre o bico da torneira. Também iniciamos o desenvolvimento de uma técnica para o controle de caos. Verificamos a existência de pontos de sela em vários atratores experimentais e, com a aplicação de dinâmica simbólica, observamos tangências homoclínicas associadas ao aparecimento de atratores de Hénon e bifurcações homoclínicas. Utilizando métodos de caracterização topológica, estabelecemos duas rotas para o caos envolvendo tangências homoclínicas, e mostramos que o súbito desaparecimento de um atrator caótico, em altas taxas de vazão, é devido a uma \"chaotic blue sky catastrophe\", apenas observada anteriormente num modelo de equações usadas por Van der Pol para simular a dinâmica cardíaca. / We assembled an experimental apparatus to study the dynamical complex behavior of water drop formation in a nipple faucet. We developed a closed hydrodynamic circuitry, and an automated acquisition data system, which also controls the faucet (a needle valve) opening. We have used as a control parameter the dripping rate set up by the faucet opening. For each dripping rate, the data are interdrop time series {Tn} between two successive drops. With the help of bifurcation diagrams, and reconstructed phase spaces in first return maps Tn+I x Tn, we were able to observe period doubling, Hopf bifurcation, interior and boundary crises, intermittent behaviors, and quasiperiodic movements. An anti-control of chaos was applied by perturbing a stable fixed point with pulses of compressed air on the nipple faucet. We also started the development of a technique to apply the control of chaos. The occurrence of saddle points was verified in some experimental attractors. By applying symbolic dynamics, we were able to observe homoclinic tangencies associated with the appearence of Hénon-like attractors and homoclinic bifurcations. By means of topological characterization, we established two routes to chaos related to homoclinic tangencies. We also observed, at high dripping rates, a sudden disappearance of a chaotic attractor due to a \"chaotic blue sky catastrophe\", just seen in a Van der Pol model used to simulate cardiac dynamics.

O planejamento da operação energética de sistemas hidrotérmicos de potência tem como objetivo determinar a participação de usinas termoelétricas e hidrelétricas de forma a garantir o suprimento desta energia ao menor custo operacional possível, dentro de restrições técnicas. Alguns fatores tornam a solução desse problema bastante complexa destacando a não linearidade, presente na equação de geração hidráulica; a não separabilidade espacial, devido ao fato da decisão de quanto gerar em uma usina interferir em outra usina do sistema; a separabilidade temporal aditiva, devido a interferência de uma decisão atual em uma decisão futura e, como no caso brasileiro, de grande porte. O objetivo deste trabalho é apresentar uma nova abordagem para o planejamento da operação de sistemas hidrotérmicos de potência, com Programação Dinâmica, de forma que as usinas hidrelétricas possam ser representadas e otimizadas individualmente, completamente desacopladas. Essa aplicação é possível através da utilização de uma função objetivo modificada, considerando-se não apenas os custos, mas também os dados de afluências das usinas imediatamente a jusante. O modelo proposto, como função objetivo modificada, foi aplicado em cascatas de usinas hidrelétricas brasileiras, cujos resultados puderam ser comparados com os comportamentos ótimos obtidos de forma direta, ou seja, com a otimização de todo o sistema em conjunto. Com base nos resultados obtidos, pôde-se observar o bom comportamento do modelo implementado. / The energetic operation planning in hydrothermal power systems aims to determine the participation of hydroelectric power plants and thermoelectric power plants to ensure the supply of electricity at the lowest operating cost possible, between technical constraints. Some factors make the solution of this problem quite complex highlighting the nonlinearity, present in the equation of hydraulic generation, non-spatial separability, due the fact that the decision of the amount of energy generated in a plant interfere in another plants of the system, the additive time separability, because of the interference from an actual decision on a future decision and, as in the Brazilian case, large. The proposal of this work is to present a new approach to hydrothermal power systems, with Dynamic Programming, so that hydroelectric power plants can be represented and optimized individually, completely decoupled. This application is possible through the use of a modified objective function, considering not only costs but also the inflow data immediately on the downstream plants. The proposed model was applied in a Brazilian cascade of hydroelectric plants, comparing the optimal behavior obtained directly, by optimizing the whole system together and the new technique behavior. Based on the results obtained, we can observe the good results of the implemented model.

Depuis deux décennies, les nombreux développements des techniques de radiothérapie par modulation d’intensité (RCMI) ont permis de mieux conformer la dose au volume cible et ainsi, d’augmenter les taux de réussite des traitements des cancers. Ces techniques ont souvent l’avantage de réduire la dose aux organes à risque proches de la zone traitée, mais elles ont l’inconvénient d’apporter un niveau de dose périphérique plus important que les techniques basiques sans modulation d’intensité. Dans ce contexte, l’augmentation du taux de survie des patients qui en résulte, accroît également la probabilité de manifestation d’effets iatrogènes dus aux doses périphériques (tels que les cancers secondaires). Aujourd’hui, la dose périphérique n’est pas considérée lors de la planification du traitement et il n’existe aucun outil numérique fiable pour sa prédiction. Il devient cependant indispensable de prendre en compte le dépôt de dose périphérique lors de la planification du traitement, notamment dans les cas pédiatriques. Cette étude doctorale a permis la réalisation de plusieurs étapes du développement d’un outil numérique, précis et rapide, de prédiction de la dose hors-champ fondé sur le code Monte Carlo PENELOPE. Dans cet objectif, nous avons démontré la capacité du code PENELOPE à estimer la dose périphérique en comparant ses résultats avec des mesures de référence réalisées à partir de deux configurations expérimentales (métrologique et pré-clinique). Ces travaux expérimentaux ont notamment permis la mise en place d’un protocole d’utilisation des dosimètres OSL pour la mesure des faibles doses. En parallèle, nous avons pu mettre en évidence la convergence lente et rédhibitoire du calcul en vue d’une utilisation clinique. Par conséquent nous avons réalisé un travail d’accélération du code en implémentant une nouvelle technique de réduction de variance appelée transport pseudo-déterministe spécifiquement dédiée à l’amélioration de la convergence dans des zones lointaines du faisceau principal. Ces travaux ont permis d’améliorer l’efficacité des estimations dans les deux configurations de validation définies (gain d’un facteur 20) pour atteindre des temps de calcul raisonnables pour une application clinique. Des travaux d’optimisation du code restent à entreprendre de façon à améliorer encore la convergence de l’outil pour ensuite en envisager une utilisation clinique. / Over the last two decades, many technical developments have been achieved on intensity modulated radiotherapy (IMRT) and allow a better conformation of the dose to the tumor and consequently increase the success of cancer treatments. These techniques often reduce the dose to organs at risk close to the target volume; nevertheless they increase peripheral dose levels. In this situation, the rising of the survival rate also increases the probability of secondary effects expression caused by peripheral dose deposition (second cancers for instance). Nowadays, the peripheral dose is not taken into account during the treatment planification and no reliable prediction tool exists. However it becomes crucial to consider the peripheral dose during the planification, especially for pediatric cases. Many steps of the development of an accurate and fast Monte Carlo out-of-field dose prediction tool based on the PENELOPE code have been achieved during this PhD work. To this end, we demonstrated the ability of the PENELOPE code to estimate the peripheral dose by comparing its results with reference measurements performed on two experimental configurations (metrological and pre-clinical). During this experimental work, we defined a protocol for low doses measurement with OSL dosimeters. In parallel, we highlighted the slow convergence of the code for clinical use. Consequently, we accelerated the code by implementing a new variance reduction technique called pseudo-deterministic transport which is specifically with the objective of improving calculations in areas far away from the beam. This step improved the efficiency of the peripheral doses estimation in both validation configurations (by a factor of 20) in order to reach reasonable computing times for clinical application. Optimization works must be realized in order improve the convergence of our tool and consider a final clinical use.

In this thesis, the effect of pitch based optimisation was explored for a 80 turbine wind farm. Using a modified Jensen wake model and the Particle Swarm Optimisation (PSO) model, a pitch optimisation strategy was created for the dominant turbulence and atmospheric condition for the wind farm. As the wake model was based on the FLORIS model developed by P.M.O Gebraad et. al., the wake and power model was compared with the FLORIS model and a -0.090% difference was found. To determine the dynamic predictive capability of the wake model, measurement values across a 10 minute period for a 19 wind turbine array were used and the wake model under predicted the power production by 17.55%. Despite its poor dynamic predictive capability, the wake model was shown to accurately match the AEP production of the wind farm when compared to a CFD simulation done in FarmFlow and only gave a 3.10% over-prediction. When the optimisation model was applied with 150 iterations and particles, the AEP production of the wind farm increased by 0.1052%, proving that the pitch optimisation method works for the examined wind farm. When the iterations and particles used for the optimisation was increased to 250, the power improvement between optimised results improved by 0.1144% at a 222.5% increase in computational time, suggesting that the solution has yet to fully converge. While the solutions did not fully converge, they converged sufficiently and an increase in iterations gave diminishing results. From the results, the pitch optimisation model was found to give a significant increase in power production, especially in wake intensive wind directions. However, the dynamic predictive capabilities will have be improved upon before the control strategy can be applied to an operational wind farm.

<p>Modern semiconductor design methods makes it possible to design increasingly complex system-on-a-chips (SOCs). Testing such SOCs becomes highly expensive due to the rapidly increasing test data volumes with longer test times as a result. Several approaches exist to compress the test stimuli and where hardware is added for decompression. This master’s thesis presents a test data compression method based on a modified facsimile code. An embedded processor on the SOC is used to decompress and apply the data to the cores of the SOC. The use of already existing hardware reduces the need of additional hardware. </p><p>Test data may be rearranged in some manners which will affect the compression ratio. Several modifications are discussed and tested. To be realistic a decompressing algorithm has to be able to run on a system with limited resources. With an assembler implementation it is shown that the proposed method can be effectively realized in such environments. Experimental results where the proposed method is applied to benchmark circuits show that the method compares well with similar methods. </p><p>A method of including the response vector is also presented. This approach makes it possible to abort a test as soon as an error is discovered, still compressing the data used. To correctly compare the test response with the expected one the data needs to include don’t care bits. The technique uses a mask vector to mark the don’t care bits. The test vector, response vector and mask vector is merged in four different ways to find the most optimal way.</p>

In this dissertation, we study data collection and its achievable network capacity in Wireless Sensor Networks (WSNs). Firstly, we investigate the data collection issue in dual-radio multi-channel WSNs under the protocol interference model. We propose a multi-path scheduling algorithm for snapshot data collection, which has a tighter capacity bound than the existing best result, and a novel continuous data collection algorithm with comprehensive capacity analysis. Secondly, considering most existing works for the capacity issue are based on the ideal deterministic network model, we study the data collection problem for practical probabilistic WSNs. We design a cell-based path scheduling algorithm and a zone-based pipeline scheduling algorithm for snapshot and continuous data collection in probabilistic WSNs, respectively. By analysis, we show that the proposed algorithms have competitive capacity performance compared with existing works. Thirdly, most of the existing works studying the data collection capacity issue are for centralized synchronous WSNs. However, wireless networks are more likely to be distributed asynchronous systems. Therefore, we investigate the achievable data collection capacity of realistic distributed asynchronous WSNs and propose a data collection algorithm with fairness consideration. Theoretical analysis of the proposed algorithm shows that its achievable network capacity is order-optimal as centralized and synchronized algorithms do and independent of network size. Finally, for completeness, we study the data aggregation issue for realistic probabilistic WSNs. We propose order-optimal scheduling algorithms for snapshot and continuous data aggregation under the physical interference model.

In classical network information theory, information packets are treated as commodities, and the nodes of the network are only allowed to duplicate and forward the packets. The new paradigm of network coding, which was introduced by Ahlswede et al., states that if the nodes are permitted to combine the information packets and forward a function of them, the throughput of the network can dramatically increase. In this dissertation we focused on the design and analysis of low complexity network coding schemes for different topologies of wired and wireless networks. In the first part we studied the routing capacity of wired networks. We provided a description of the routing capacity region in terms of a finite set of linear inequalities. We next used this result to study the routing capacity region of undirected ring networks for two multimessage scenarios. Finally, we used new network coding bounds to prove the optimality of routing schemes in these two scenarios. In the second part, we studied node-constrained line and star networks. We derived the multiple multicast capacity region of node-constrained line networks based on a low complexity binary linear coding scheme. For star networks, we examined the multiple unicast problem and offered a linear coding scheme. Then we made a connection between the network coding in a node-constrained star network and the problem of index coding with side information. In the third part, we studied the linear deterministic model of relay networks (LDRN). We focused on a unicast session and derived a simple capacity-achieving transmission scheme. We obtained our scheme by a connection to the submodular flow problem through the application of tools from matroid theory and submodular optimization theory. We also offered polynomial-time algorithms for calculating the capacity of the network and the optimal coding scheme. In the final part, we considered the multicasting problem in an LDRN and proposed a new way to construct a coding scheme. Our construction is based on the notion of flow for a unicast session in the third part of this dissertation. We presented randomized and deterministic polynomial-time versions of our algorithm.

Higher Order Statistics (HOS) and Second Order Statistics (SOS) approaches have certain advantages and disadvantages in signal processing applications. HOS approach provides more statistical information for non-Gaussian signals. On the other hand, SOS approach is more robust to the estimation errors than the HOS approach, especially when the number of observations is small. In this thesis, HOS and SOS approaches are jointly used in order to take advantage of both methods. In this respect, the joint use of HOS and SOS approaches are introduced for online calibration of sensor arrays with arbitrary geometries. Three different problems in online array calibration are considered and new algorithms for each of these problems are proposed. In the first problem, the positions of the randomly deployed sensors are completely unknown except the two reference sensors and HOS and SOS approaches are used iteratively for the joint Direction of Arrival (DOA) and sensor position estimation. Iterative HOS-SOS algorithm (IHOSS) solves the ambiguity problem in sensor position estimation by observing the source signals at least in two different frequencies and hence it is applicable for wideband signals. The conditions on these frequencies are presented. IHOSS is the first algorithm in the literature which finds the DOA and sensor position estimations in case of randomly deployed sensors with unknown coordinates. In the second problem, narrowband signals are considered and it is assumed that the nominal sensor positions are known. Modified IHOSS (MIHOSS) algorithm uses the nominal sensor positions to solve the ambiguity problem in sensor position estimation. This algorithm can handle both small and large errors in sensor positions. The upper bound of perturbations for unambiguous sensor position estimation is presented. In the last problem, an online array calibration method is proposed for sensor arrays where the sensors have unknown gain/phase mismatches and mutual coupling coefficients. In this case, sensor positions are assumed to be known. The mutual coupling matrix is unstructured. The two reference sensors are assumed to be perfectly calibrated. IHOSS algorithm is adapted for online calibration and parameter estimation, and hence CIHOSS algorithm is obtained. While CIHOSS originates from IHOSS, it is fundamentally different in many aspects. CIHOSS uses multiple virtual ESPRIT structures and employs an alignment technique to order the elements of rows of the actual array steering matrix. In this thesis, a new cumulant matrix estimation technique is proposed for the HOS approach by converting the multi-source problem into a single source one. The proposed algorithms perform well even in the case of correlated source signals due to the effectiveness of the proposed cumulant matrix estimate. The iterative procedure in all the proposed algorithms is guaranteed to converge. Closed form expressions are derived for the deterministic Cram&acute / er-Rao bound (CRB) for DOA and unknown calibration parameters for non-circular complex Gaussian noise with unknown covariance matrix. Simulation results show that the performances of the proposed methods approach to the CRB for both DOA and unknown calibration parameter estimations for high SNR.

In this thesis we address the problem of designing codes for specific applications. To do so we make use of the relationship between clusters and codes. Designing a block code over any finite dimensional space may be thought of as forming the corresponding number of clusters over the particular dimensional space. In literature we have a number of algorithms available for clustering. We have examined the performance of a number of such algorithms, such as Linde-Buzo-Gray, Simulated Annealing, Simulated Annealing with Linde-Buzo-Gray, Deterministic Annealing, etc, for design of codes. But all these algorithms make use of the Eucledian squared error distance measure for clustering. This distance measure does not match with the distance measure of interest in the error correcting scenario, namely, Hamming distance. Consequently we have developed an algorithm that can be used for clustering with Hamming distance as the distance measure. Also, it has been observed that stochastic algorithms, such as Simulated Annealing fail to produce optimum codes due to very slow convergence near the end. As a remedy, we have proposed a modification based on the code structure, for such algorithms for code design which makes it possible to converge to the optimum codes.

The implicit assumption in conventional inventory models is that the stored products maintain the same utility forever, i.e., they can be stored for an infinite period of time without losing their value or characteristics. However, generally speaking, almost all products experience some sort of deterioration over time. Some products have very small deterioration rates, and henceforth the effect of such deterioration can be neglected. Some products may be subject to significant rates of deterioration. Fruits, vegetables, drugs, alcohol and radioactive materials are examples that can experience significant deterioration during storage. Therefore the effect of deterioration must be explicitly taken into account in developing inventory models for such products. In most existing deteriorating inventory models, time is treated as a continuous variable, which is not exactly the case in practice. In real-life problems time factor is always measured on a discrete scale only, i.e. in terms of complete units of days, weeks, etc. In this research, we present several discrete-in-time inventory models and identify optimal ordering policies for a single deteriorating product by minimizing the expected overall costs over the planning horizon. The various conditions have been considered, e.g. periodic review, time-varying deterioration rate, waiting-time-dependent partial backlogging, time-dependent demand, stochastic demand etc. The objective of our research is two-fold: (a) To obtain optimal order quantity and useful insights for the inventory control of a single deteriorating product over a discrete time horizon with deterministic demand, variable deterioration rates and waiting-time-dependent partial backlogging ratios; (b) To identify optimal ordering policy for a single deteriorating product over a finite horizon with stochastic demand and partial backlogging. The explicit ordering policy will be developed for some special cases. Through computational experiments and sensitivity analysis, a thorough and insightful understanding of deteriorating inventory management will be achieved.