Computer networks can be classified into two broad categories: wired networks and wireless networks, according to the hardware and software technologies used to interconnect the individual devices. Wired interconnection networks are hardware fabrics supporting communications between individual processors in highperformance computing systems (e.g., multi-computer systems and cluster systems). On the other hand, due to the rapid development of wireless technologies, wireless networks have emerged and become an indispensable part for people's lives. The integration of different wireless technologies is an effective approach to accommodate the increasing demand of the users to communicate with each other and access the Internet. This thesis aims to investigate the performance of wired interconnection networks and integrated wireless networks under the realistic working conditions. Traffic patterns have a significant impact on network performance. A number of recent measurement studies have convincingly demonstrated that the traffic generated by many real-world applications in communication networks exhibits bursty arrival nature and the message destinations are non-uniformly distributed. Analytical models for the performance evaluation of wired interconnection networks and integrated wireless networks have been widely reported. However, most of these models are developed under the simplified assumption of non-bursty Poisson process with uniformly distributed message destinations. To fill this gap, this thesis first presents an analytical model to investigate the performance of wired interconnection networks in multi-computer systems. Secondly, the analytical models for wired interconnection networks in multi-cluster systems are developed. Finally, this thesis proposes analytical models to evaluate the end-to-end delay and throughput of integrated wireless local area networks and wireless mesh networks. These models are derived when the networks are subject to bursty traffic with non-uniformly distributed message destinations which can capture the burstiness of real-world network traffic in the both temporal domain and spatial domain. Extensive simulation experiments are conducted to validate the accuracy of the analytical models. The models are then used as practical and cost-effective tools to investigate the performance of heterogeneous wired or wireless networks under the traffic patterns exhibited by real-world applications.

The world is currently experiencing major issues in the energy sector. The ever-growing human population, limited energy resources and the effect of greenhouse gas emissions have become major global concerns for the energy sector, including the electricity generation sector. This dilemma caused electricity providers to revise their generation methods and created a major need for consumers to utilise electricity more efficiently. Demand side management (DSM) is one initiative developed for consumers to efficiently utilise electricity. Due to their high electricity consumption and technical skills, mines are ideal targets for the implementation of DSM strategies. Therefore, the focus of this study was to investigate South African mines for possible implementation of DSM strategies on their compressed air networks. Compressed air networks at South African mines are relatively old and inadequately maintained. This causes inefficient distribution and use of compressed air. The study will therefore focus on reconfiguring mining compressed air networks for cost savings. Cost savings include financial savings on electricity bills, implementation costs and decreased maintenance. Through several investigations, the possibility of implementing energy savings strategies to reconfigure the compressed air networks of two South African mines was identified. Reconfiguring the networks would respectively entail interconnecting two shafts and relocating a compressor from an abandoned shaft to a fully productive shaft. Theoretical simulations were developed to determine the networks’ responses to the reconfiguration strategies. The simulations assisted in exposing the viability of implementing the reconfiguration strategies on the respective compressed air networks. Positive responses were obtained from the simulations and proposals were made to the respective mines for possible implementation. The proposed initiatives were implemented on the respective mines’ compressed air networks. After implementation of the interconnection strategy, a consecutive three-month performance assessment period commenced to prove the viability of the proposed savings. An average power saving of 1 700 kW was achieved during the performance assessment period. The proposed initiative to relocate the compressor is currently being implemented. A financial saving of approximately R8.9 million per annum was achieved by implementing the interconnection strategy. The large financial saving was due to the utilisation of the mine’s salvaged equipment. Further savings were achieved by the decreased maintenance on the mine’s compressors. Due to the successful implementation of the interconnection strategy, it is safe to state that cost savings can be achieved by reconfiguring mining compressed air networks. / MIng (Mechanical Engineering), North-West University, Potchefstroom Campus, 2014

L’évolution de la technologie microélectronique conduit à une densité d’intégration toujours plus forte des transistors. Les structures d’interconnexions en cuivre Damascène suivent cette tendance et doivent être maîtrisées en termes de fabrication, de performance et de robustesse, ces différents aspects étant intimement liés aux contraintes résiduelles et à la résistivité. Cette thèse vise à comprendre les mécanismes de génération de contraintes et identifier les différentes contributions à la résistivité de ces objets en fonction des conditions de recuit et des dimensions (de la centaine de nm à plusieurs µm). Pour ce faire, les rôles respectifs de la microstructure et des dimensions de films et de lignes de cuivre électrodéposés ont été découplés sur la base de modèles analytiques intégrants des paramètres microstructuraux et géométriques. La microstructure a été analysée principalement à partir de cartographies d’orientations cristallines réalisées par EBSD. Dans le cas des lignes de cuivre de 0.2 à 1 µm de large, les contraintes résiduelles ont été déduites de l’exploitation de nano-capteurs pivotants spécialement élaborés. Les résultats obtenus montrent qu’indépendamment de la température de recuit, l’augmentation de résistivité et de contraintes résiduelles observée vers les faibles dimensions est le fruit d’une diminution de la taille moyenne de cristallites et d’un confinement géométrique plus prononcé. En outre, l’augmentation de résistivité résulte également d’une élévation de la probabilité de réflexion des électrons aux joints de grains. Cette dernière a été associée à la réduction de la proportion de joints de grains spéciaux de cohérence atomique élevée. / The evolution of the microelectronic technology leads to a transistors integration density always stronger. The Damascene copper interconnections structures follow this tendency and must be controlled in terms of manufacturing, performance and robustness, these different aspects being intimately related to the residual stresses and resistivity. This thesis aims to understand the mechanisms of the residual stresses generation and identify the different contributions to the resistivity of these objects as a function of annealing conditions and dimensions (from about a hundred of nm to several µm). In order to do this, the respective effects of the microstructure and dimensions of electroplated copper films and lines were separated on the basis of analytical models integrating microstructural and geometrical parameters. The microstructure was principally analysed from mappings of crystalline orientations achieved by EBSD. For the copper lines of width 0.2 and 1 µm, the residual stresses were deduced from the exploitation of nano-rotating sensors specially elaborated. The results obtained show that independently of the annealing temperature, the resistivity and residual stresses increase observed toward the small dimensions arises from the diminution of the average crystallites size and the geometrical confinement more pronounced. Furthermore, the resistivity increase results also of the electrons reflection probability growth at grains boundaries. This last point was associated to the reduction of the proportion of special grains boundaries having a high atomic coherency.

More and more system operators are interested in calculating transfer capability in real-time using real-time power flow models generated from the Energy Management System (EMS). However, compared to off-line study models, EMS models usually cover only a limited portion of the interconnected system. In most situations, it is not practical to extend the EMS model to capture the impact of the external systems and therefore using an equivalent network becomes necessary. The development of equivalent circuits to represent external areas was a topic discussed over the last 50 years. Almost all of these methods require impedance information about the external area to develop the equivalent circuit. Unfortunately utilities do not have the external impedance information in the real-time. Therefore, normal industry practice is to use off-line studies to develop an equivalent circuit and use that circuit in the real-time operation without any validation. This can result in errors in the security assessment. Therefore, power industry need a method to develop or validate an equivalent circuit based on the available real-time information. This thesis work is focussed on meeting that industry need. The work on this thesis presents two new methods that can be used to generate an equivalent circuit based on the boundary conditions. This method involves calculating equivalent impedance between two areas based on the boundary stations voltages, voltage angles and power leaving the boundary stations into external areas. This thesis uses power system simulation between two areas to change the system condition to obtain different boundary bus voltages, voltage angles and power injections to generate necessary data. Regression analysis and least square method is then used to generate the equivalent circuit using these data. It is expected that system changes will provide necessary information in the real-time to generate the equivalent circuit. The proposed methodology is validated with modified three area 300 bus system as well as using Manitoba Hydro’s system. Contingency analysis, transfer level calcula-tion and PV curves analysis is used to validate the proposed method. Simulation results show that the proposed method produces adequate accuracy in comparison with detailed off-line system models. The main advantage of the proposed method as compared to other existing meth-ods such as Ward and REI is that the proposed method does not require external imped-ance information to generate the equivalent circuit. The ability to generate reasonably good equivalent circuit only using available boundary information will help utilities to generate or validate the equivalent circuit based on the current system conditions, which will intern help improve the accuracy of the security assessment / February 2017

This diploma thesis deals with the technical analysis with the emphasis on creating, testing and using of trading systems. Its objective is to find out whether it is possible for a trader to design and trade his own profitable trading system with widely accessible tools and methods. First part of the thesis concentrates among other things on the explanation of stock valuation principles, description of tested shares and time period. The second and third chapters fully describe the process of trading system development and the analysis of results of both trading systems. Last chapter is devoted to the interconnection of European stock markets, which is explored by the means of correlation analysis among different stock indexes. The correlation coefficients show a strong link of the markets and the rising level of integration of European markets.

Made available in DSpace on 2016-04-29T13:17:36Z (GMT). No. of bitstreams: 1 Jan Luiz Leonardi.pdf: 974669 bytes, checksum: ff14eae278e1093b0a83b8d1bf2d5ea9 (MD5) Previous issue date: 2012-02-10 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Insight has been interpreted in behavior analysis as the sudden resolution of a problem when two or more different repertoires, previously learned separately, interconnect in a new situation without direct training, producing original sequences of behavior. A small body of research has shown some variables which appear to contribute to the production of this phenomenon, but methodological problems, in particular the use of arbitrary objects in the problem-situation and the delay between response and reinforcer, limit its relevance. Under the light of these considerations, the objective of this research was to investigate the phenomenon of spontaneous interconnection of repertoires (or insight ) with variations in the procedures employed in the area. These were: (a) the object that should be reached by the subjects in the problem-situation was established as a reinforcing stimulus before the phases of testing and training; (b) the delay between the target response (from training and testing) and consumption of the primary reinforcer (water) was reduced by the construction of a drinking magazine that could be moved toward the subject. Two Wistar rats were used as experimental subjects. The equipment and materials were a circular chamber of 69 cm in diameter and 50 cm high, four acrylic cubes coated with black cardboard, a nylon string that drove the movement of the cube along the camera, and a buzzer. The procedure was divided in six phases: (a) training to the drinking magazine, (b) pre-test of insight, (c) training to push the cube, (d) intermediate test of insight, (e) training to climb on the cube and rise, (f) final test of insight. In the test of insight situation, the subject should reach the drinking magazine positioned at the top of the chamber by pushing an acrylic cube in its direction and climbing on the cube to drink. Water were used as reinforcement in a continuous schedule of reinforcement. The results showed that none of the subjects solved the problem in a sudden, direct and continuous way, criteria used to qualify the performance as insight . Nevertheless, the procedure used in this experiment has heuristic value as it represents an attempt to produce the phenomenon without the use of arbitrary objects (which significantly reduces the number of training sessions), and rendered the concept of functional generalization unnecessary. Finally, this research points out obstacles that the area faces in the study of insight , such as the difficulty in replicating Epstein s research and the problems generated by the use of arbitrary objects in the problem-situation, not to mention the limitations imposed by the behavioral interpretation of the phenomenon / O insight tem sido interpretado na análise do comportamento como a resolução súbita de um problema quando dois ou mais repertórios diferentes, previamente aprendidos em separado, se interconectam sem treino direto em uma nova situação, produzindo sequências originais de comportamento. Um pequeno conjunto de pesquisas evidenciou algumas variáveis que parecem contribuir para a produção desse fenômeno, mas problemas metodológicos, em especial o uso de objetos arbitrários na situação-problema e o atraso entre resposta e reforçador, limitam sua relevância. Em vista disso, o objetivo desta pesquisa foi investigar o fenômeno de interconexão espontânea de repertórios (ou insight ) implementando algumas variações nos procedimentos empregados pela área. Foram elas: (a) o objeto que deveria ser alcançado pelos sujeitos na situação-problema foi estabelecido como estímulo reforçador antes das fases de teste e de treino; (b) o atraso entre a resposta-alvo (dos treinos e dos testes) e o consumo do reforçador primário (água) foi reduzido por meio da construção de um bebedouro que podia ser deslocado até o sujeito. Dois ratos Wistar foram utilizados como sujeitos experimentais. Os equipamentos e materiais foram uma câmara circular de 69 cm de diâmetro e 50 cm de altura, quatro cubos de acrílico revestidos com cartolina preta, um fio de náilon que guiava o movimento do cubo ao longo da câmara e um buzzer. O procedimento foi dividido em seis fases: (a) treino ao bebedouro; (b) pré-teste de insight; (c) treino de empurrar o cubo; (d) teste intermediário de insight; (e) treino de subir no cubo e erguer-se; (f) teste final de insight. Na situação de teste de insight, o sujeito deveria alcançar o bebedouro, posicionado no alto da câmara, empurrando um cubo de acrílico em direção a ele e subindo no cubo para beber. Gotas de água foram utilizadas como reforçador em esquema de reforçamento contínuo. Os resultados demonstraram que nenhum dos sujeitos resolveu o problema de forma súbita, direta e contínua, critérios empregados para qualificar o desempenho como insight . Apesar disso, o procedimento empregado neste experimento tem valor heurístico, na medida em que representa uma tentativa de produzir o fenômeno sem o uso de objetos arbitrários (o que reduz significativamente o número de sessões de treino necessárias), além de tornar desnecessário o conceito de generalização funcional. Por fim, são apontados alguns obstáculos que a área enfrenta no estudo do insight , como a dificuldade em replicar as pesquisas de Epstein e os problemas gerados pelo uso de objetos arbitrários na situação-problema, além de limitações impostas pela própria interpretação comportamental do fenômeno

Made available in DSpace on 2016-04-29T13:18:09Z (GMT). No. of bitstreams: 1 Juliana de Souza Ferreira.pdf: 3589110 bytes, checksum: c8e5dedeebec46a11c8bc2a721d030a0 (MD5) Previous issue date: 2008-05-09 / The present study replies Epstein, Kirshnit, Lanza and Rubin s (1984) research in which pigeons solved a problem by pushing a box beneath an unattainable target, climbing this box and pecking the target. It was investigated if new behaviors emerge from the interconnection of two trained repertoires with rats and if changes in the sequence order in which these two repertoires was taught would modify subjects performance in the final test. The research was carried out with 3 rats. The equipments were a cylindrical chamber with 69 cm in diameter, one triangle and a green led used as targets, and an acrylic box. Two skills were trained, with the 3 subjects, separated. 1 subject has learned 1) to climb a box and pull a triangle and 2) to push a box toward a green light. This training order associated with individual distinction implicated in a longer training of both skills. The other 2 subjects have learned 1) to push a box toward a green light and 2) to climb a box and pull a triangle. This training order implicated in a shorter training of both skills. Finally, an insight test was carried out: the triangle was fixed out of reach and the box was placed in the other side of the chamber. The subjects should push the box beneath the target, climb the box and pull the triangle. The results showed that one subject that was submitted to a longer training, solved the problem with suddenness, directness and continuousness and satisfied the insight s criterion; another subject that was submitted to a shorter training, solved the problem, but between pushing the box and climbing it, a 2 minutes interval has passed over, the response chain didn t occur continuously; and a third subject didn t solve the problem even when the training was similar to the subject that solved the problem discontinuously / O presente estudo replicou a pesquisa de Epstein, Kirshnit, Lanza e Rubin (1984) no qual pombos conseguiram resolver um problema ao empurrar uma caixa até um alvo que estava fora do alcance, subir na caixa e bicar o alvo. Foi investigado se comportamentos novos surgem por meio da interconexão de dois repertórios previamente treinados em ratos e se mudanças na ordem do treino desses dois repertórios alteram o desempenho do sujeito no teste. A pesquisa foi realizada com 3 ratos. O equipamento utilizado foi uma câmara circular de 69 cm de diâmetro, um triângulo, um led verde (luz) e uma caixa de acrílico. Foram treinadas duas habilidades, com os 3 sujeitos, separadamente. 1 sujeito aprendeu a: 1) subir na caixa e puxar o triângulo e 2) empurrar a caixa direcionadamente a uma luz. Esta ordem do treino associada às diferenças individuais implicou em um maior número de sessões de treino das 2 habilidades para este sujeito. Os outros 2 sujeitos aprenderam a: 1) empurrar a caixa direcionadamente a uma luz e 2) subir na caixa e puxar o triângulo. Esta ordem do treino implicou em um menor número de sessões de treino das 2 habilidades para os 2 sujeitos. Por fim, foi realizado um teste de insight: o triângulo foi colocado fora de alcance e a caixa foi colocada do outro lado da câmara. O sujeito deveria empurrar a caixa até debaixo do triângulo, subir na caixa e puxá-lo. Os resultados mostraram que o sujeito que teve um treino mais longo, resolveu o problema de forma direcionada, contínua e súbita e satisfez os critérios para considerar a solução um insight; um outro sujeito que passou pelo treino mais curto resolveu o problema, no entanto, entre a resposta de empurrar a caixa e as respostas de subir na caixa e puxar o triângulo transcorreu um intervalo de 2 minutos, a cadeia de respostas não ocorreu de forma contínua; e um terceiro sujeito não resolveu o problema, apesar de ter passado por um treino parecido ao do sujeito que resolveu de forma descontínua

Cette thèse, centrée en Automatique, porte sur l’approximation de systèmes linéaires de dimension infinie en une dimension d’espace d’un point de vue entrée-sortie par une classe de systèmes à retards, et sur l’analyse de stabilité de cette classe d’approximation. Afin de conserver des propriétés du système de dimension infinie par interconnexion (stabilité, performance entrée-sortie), l’approximation est définie sur la topologie du graphe. Classiquement, les méthodes proposées par l’approximation conduisent à des modèles de dimension finie. Cependant, sur cette topologie du graphe, une telle approximation régie par des équations à paramètres localisés peut ne pas exister. On propose donc d’étendre cette classe d’approximation en y incluant l’opérateur retard. Nous obtenons alors un modèle d’équations différentielles couplées à des équations aux différences. L’existence et la mise en œuvre numérique de cette classe d’approximation et ses propriétés de réalisation d’état sont étudiées. Le deuxième enjeu de nos recherches s’est focalisé sur l’analyse de stabilité de cette classe d’approximation, par le biais de l’approche de Lyapunov-Krasovskii. Cette approche consiste à exploiter des conditions de stabilité sous forme d’un problème l’optimisation convexe. Cette analyse de stabilité est étendue au cas des systèmes avec paramètres incertains et des retards variants dans le temps pour la sous classe des équations aux différences. Nous développons également une estimation de la décroissance de la solution en vue de l’analyse de performance. L’analyse du conservatisme de la méthodologie proposée a été étudié. / This thesis addresses the approximation of infinite-dimensional linear systems for one-dimension in space, with the input-output approach, by a class of delay systems, and the stability analysis of this class of approximation. In order to preserve the properties of the original properties by interconnection, such as stability and/or input-output performance, the approximation is defined within the graph topology framework. In general, the methods proposed in the literature lead, by approximation, to finite-dimensional models. However, in this topology, such an approximation by lumped plants may not exist. It seems natural to generalize this approximation class by including a delay operator. We then obtain an approximation in the graph topology governed by a model of coupled differential-difference equations. The existence of this class of approximation and the properties of state-realization are studied. A constructive numerical method is proposed for this approximation. After the description of this class of approximation, we investigate stability of this class of operators, by the Lyapunov-Krasovskii approach. This approach consists to involve stability conditions of the form of a convex optimization problem. This stability analysis is extended to the case of systems with uncertain parameters and time-varying delays. We also propose estimates of the decay rate of the solution for the performance analysis. The analysis of the conservatism of the proposed methodology has been studied.

Os Sistemas Supervisórios (SS) executam diversas funções vitais em um processo automatizado e também operam como interface homem-máquina. Os mesmos recebem informações de dispositivos como Controladores Lógicos Programáveis (CLP), inversores de freqüência, etc, e ao mesmo tempo enviam parâmetros de controle fornecidos pelos operadores do processo aos equipamentos de controle. Na operação de SS, a atuação correta e a experiência dos operadores humanos é portanto também vital no controle do processo automatizado. Em recente trabalho(Andrade-2001) de pesquisa na Escola Politécnica da USP foi desenvolvido um Sistema Especialista para operar em tempo real com Sistemas Supervisórios para auxiliar na tomada de decisão dos operadores do sistema. Ao longo do tempo de operação de uma planta automatizada ocorrem novas situações que passam a compor os novos cenários do sistema e também contribuem para o aumento do conhecimento e da experiência dos operadores humanos. Assim sendo, Sistemas Especialistas constantemente devem ser atualizados com novas regras para atender às novas demandas da planta automatizada. Este trabalho de pesquisa apresenta os resultados obtidos com o Sistema Especialista desenvolvido para operar em tempo real com Sistemas Supervisórios, como também o andamento das pesquisas no campo de aprendizado de máquina e mineração de dados com o objetivo de desenvolver e de habilitar Sistemas Especialistas que operam em tempo real com a capacidade de aprender conforme ocorrem eventos durante o funcionamento de uma planta industrial automatizada. / Supervision systems ( SS ) perform diverse vital functions in an automated process and also operate as a man machine interface. The SS receive informations from devices as programmable logical Controllers (PLC), frequency inverters , etc, and at the same time send parameters of control supplied by the process users to the control equipment. In the SS operation, the correct actions and the experience of the human users are therefore vital in the control of the automated process. In a recent research work(Andrade-2001) of the USP Polytechnic School an Expert System was developed to interact in real time with a SS aiming to help in the decision-making process of the system users. During the operation time of an automated plant new situations occur, which come to compose the new system settings and also contribute to the increasing of the human users\' knowledge and experience. Thus, Expert Systems should be constantly brought up to date with these new rules for attend the new demands of the automated plant. This research work shows the results obtained by the expert system, developed to operate in real time with the supervisory systems. It also shows the course of the researches in the fields of machine learning and data mining with the objective of developing and enabling Expert Systems that operate in real time with the capacity to learn events as they happen during the operation of an automated industrial plant.

Made available in DSpace on 2017-07-10T17:11:46Z (GMT). No. of bitstreams: 1 DISSERTACAO Katiani Pereira2.pdf: 3768192 bytes, checksum: daffa714558affb1cf5db22181b4a527 (MD5) Previous issue date: 2014-02-21 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The development of a mathematical optimization model for the allocation of switching devices and interconnecting branches in electric power distribution systems is presented. This model aims to minimize the cost of load unserved energy and investment costs. The main constraints are continuity indicators, like DEC and FEC and technical constraints from the power flow of the system. The model is a mixed integer nonlinear programming (MINLP) which proposes the building of new interconnection lines between the main branch and lateral extensions of the same feeder. The objectives of this model are to improve system reliability and minimize energy costs, the construction of the Pareto curve by varying weighting factors contained in the objective function is presented in order to find a set of solutions to be presented to the decision maker of a distribution utility. To prove the behavior of the model, a genetic algorithm was used as solution technique in several test systems. / No presente trabalho é apresentado o desenvolvimento de um modelo matemático de otimização para a alocação de chaves seccionadoras e ramos de interconexão em sistemas de distribuição de energia elétrica. Este modelo tem como objetivo minimizar os custos de energia não suprida e os custos de investimento, na alocação das chaves seccionadoras e ramais de interconexão. As principais restrições são os indicadores de continuidade DEC e FEC e o fluxo de potência do sistema, sendo este um problema de programação não linear inteiro misto (PNLIM), tendo como diferencial propor que os ramais de interconexão possam ser construídos entre o tronco principal e os ramais laterais de um mesmo alimentador. Os objetivos desse modelo são melhorar a confiabilidade do sistema e minimizar os custos de energia, desta forma apresenta-se a construção da Curva de Pareto através da variação dos fatores de peso contidos na função objetivo, com a obtenção desta curva foi possível encontrar um conjunto de soluções para que o tomador de decisões de uma empresa, possa escolher qual dessas soluções se enquadra melhor na sua proposta de trabalho. Utiliza-se o algoritmo genético como técnica de solução em vários sistemas testes para avaliar o comportamento do modelo matemático.