Thesis (Ph.D.) - Carleton University, 2007. / Includes bibliographical references (p. 234-248). Also available in electronic format on the Internet.

Thesis (M.Sc.)--York University, 2004. Graduate Programme in Computer Science and Engineering. / Typescript. Includes bibliographical references (leaves 76-79). Also available on the Internet. MODE OF ACCESS via web browser by entering the following URL: http://gateway.proquest.com/openurl?url%5Fver=Z39.88-2004&res%5Fdat=xri:pqdiss &rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&rft_dat=xri:pqdiss:MR11842

Ζούμε μια μεταβατική εποχή, που καθορίζεται από τις τεράστιες τεχνολογικές εξελίξεις, τη διεθνοποίηση των αγορών και της γνώσης και την ευκολότερη επικοινωνία των κατοίκων ολόκληρου του πλανήτη σε όλους τους τομείς δραστηριοτήτων τους. Όροι, όπως παγκοσμιοποίηση, συγχωνεύσεις εταιριών, αυξομειώσεις στα μεγέθη των επιχειρήσεων, τεχνολογική πρόοδος, περιγράφουν ποικίλες δυνάμεις στην αγορά εργασίας, οι οποίες μεταβάλλουν τους παραδοσιακούς κανόνες που επί χρόνια έχουν ρυθμίσει την επαγγελματική μας ζωή. Τα νέα τεχνολογικά μέσα μεταβάλλουν ριζικά τον τρόπο διάδοσης των επιστημονικών γνώσεων και των πληροφοριών, επηρεάζοντας, έτσι, και την οργάνωση της εκπαίδευσης, της εργασίας, της παραγωγής. Η εργασία αυτή περιγράφει την παγκοσμιοποίηση στον επιχειρηματικό χώρο και την παρουσιάζει σαν την αιτία για τις αλλαγές στη δομή και τον τρόπο λειτουργίας επιχειρήσεων και οργανισμών, στην προσπάθειά τους να ανταπεξέλθουν στις απαιτήσεις του ανταγωνισμού. Προκύπτει λοιπόν το θέμα της Διοίκησης Αλλαγής, ένας κλάδος σύγχρονος, που αποτελεί αντικείμενο μελέτης επιστημόνων και ειδικών σε ολόκληρο τον κόσμο. Η εργασία αυτή αποτελεί το πρώτο βήμα προσέγγισης του κλάδου της Διοίκησης Αλλαγής μέσα από τεχνικές Πολυκριτήριας Ανάλυσης και Υπολογιστικής Νοημοσύνης, έχοντας σαν στόχο να προτείνει μια μεθοδολογία έρευνας και μελέτης, η οποία θα οδηγήσει στη δημιουργία μοντέλων βελτιστοποίησης διαδικασίας αλλαγής στις επιχειρήσεις. / We are living a transitional period, determined by the enormous technological developments, the globalization of markets and knowledge, and facilitate communication among people throughout the world in all areas of activity. Terms such as globalization, corporate mergers, changes in the size of business,technological progress, describe various forces in the labor market, which change the traditional rules that for years have set up our professional lives. The new technology means fundamentally change the way the dissemination of scientific knowledge and information, impacting so, the organization of education, work and production. This paper describes globalization in business and present it as the reason for the changes in the structure and operation of enterprises and organizations in their efforts to meet the demands of competition. It followed the theme of Change Management, a modern industry, which is the subject of study and individual scientists around the world.This work is the first step approach to Change Management industry through technical Multicriteria Analysis and Computational Intelligence Aiming to propose a research methodology and design, which will create optimization models change process in companies.

Στην εργασία που ακολουθεί μελετήθηκε η χρήση μεθόδων της υπολογιστικής νοημοσύνης στην πρόβλεψη της κίνησης της ισοτιμίας νομισμάτων. Για να γίνει αυτό αναπτύχθηκε ένας υβριδικός αλγόριθμος που χρησιμοποιεί νευρωνικά δύκτια και γενετικούς αλγόριθμους. Στο Πρώτο Κεφάλαιο παρουσιάζεται η ϑεωρία των νευρωνικών δικτύων, οι αρχιτεκτονικές και οι μέθοδοι εκπαίδευσής τους. Επιπλέον παρουσιάζονται οι γενετικοί αλγόριθμοι και ο γενικός τρόπος λειτουργίας τους. Στο Δεύτερο Κεφάλαιο εξετάζεται το πρόβλημα της πρόγνωσης, από την σκοπιά των νευρωνικών δικτύων, καθώς και η προβλεψιμότητα των οικονομικών χρονοσειρών. Επιπλέον παρουσιάζονται υβριδικά συστήματα που χρησιμοποιούνται για την πρόβλεψη χρονοσειρών και επεξηγείται ο τρόπος λειτουργίας του αλγορίθμου που αναπτύχθηκε εδώ. Επιπλέον παρατίθενται τα αποτελέσματα της χρήσης του λογισμικού που αναπτύχθηκε, στην πρόγνωση της μεταβολής της ισοτιμίας νομισμάτων. Στο Παράρτημα παρέχεται ο πλήρης κώδικας που αναπτύχθηκε σε MATLAB. / --

[pt] A energia é assunto estratégico para a grande maioria dos países e indústrias no mundo. O consumo atual energético é de 138,32 TWh por ano e é previsto um aumento de 44% até o ano de 2030 o que demonstra um mercado em expansão. Porém, a sociedade atual exige soluções energéticas que causem o menor impacto ambiental possível, colocando em dúvida o uso das fontes de energia utilizadas atualmente. O uso da energia solar é uma alternativa para auxiliar no atendimento da futura demanda de energia. O seu principal entrave é o custo de produção de energia ser superior as fontes de energia atuais, principalmente o petróleo. Contudo nos últimos 10 anos foi verificado um crescimento exponencial na quantidade de módulos fotovoltaicos instalados em todo mundo. Nesse trabalho é realizado um estudo sobre célula solares com poços quânticos. O uso de poços quânticos já foi apontado como ferramenta para aumentar a eficiência de células fotovoltaicas. O objetivo é descrever uma metodologia baseada em algoritmos genéticos para projeto e análise desse tipo de dispositivo e estabelecer diretivas para se construir uma célula otimizada utilizando esta tecnologia. Os resultados obtidos estão de acordo com dados experimentais, demonstram a capacidade dos poços quânticos em aumentar a eficiência de uma célula e fornecem uma ferramenta tecnológica que espera-se contribuir para o desenvolvimento do país no setor energético. / [en] The energy is a strategical issue for the great majority of the countries and industries in the world. The current world energy consumption is of 138,32 TWh per year and is foreseen an increase of 44% until the year of 2030 which demonstrates a market in expansion. However, the society demands energy solutions that cause as least ambient impact as possible, putting in doubt the use of the current technologies of power plants. The utilization of solar energy is an alternative to assist in the attendance of the future demand of energy. Its main impediment is the superior cost of energy production in comparison with the current power plants, mainly the oil based ones. However in last the 10 years an exponential growth in the amount of installed photovoltaics modules worldwide was verified. In this work a study on solar cell with quantum wells is carried through. The use of quantum wells already was pointed as tool to increase the efficiency of photovoltaics cells. The objective is to describe a methodology based on genetic algorithms for project and analysis of this type of device and to establish directive to construct an optimized cell using this technology. The results are in accordance with experimental data, that demonstrates the capacity of the quantum wells in increasing the efficiency of a cell and supply a technological tool that expects to contribute for the development of the country in the energy sector.

Emotions and sentiments play a crucial role in our everyday lives. They aid decision-making, learning, communication, and situation awareness in human-centric environments. Over the past two decades, researchers in artificial intelligence have been attempting to endow machines with cognitive capabilities to recognize, infer, interpret and express emotions and sentiments. All such efforts can be attributed to affective computing, an interdisciplinary field spanning computer science, psychology, social sciences and cognitive science. Sentiment analysis and emotion recognition has also become a new trend in social media, avidly helping users understand opinions being expressed on different platforms in the web. In this thesis, we focus on developing novel methods for text-based sentiment analysis. As an application of the developed methods, we employ them to improve multimodal polarity detection and emotion recognition. Specifically, we develop innovative text and visual-based sentiment-analysis engines and use them to improve the performance of multimodal sentiment analysis. We begin by discussing challenges involved in both text-based and multimodal sentiment analysis. Next, we present a number of novel techniques to address these challenges. In particular, in the context of concept-based sentiment analysis, a paradigm gaining increasing interest recently, it is important to identify concepts in text; accordingly, we design a syntaxbased concept-extraction engine. We then exploit the extracted concepts to develop conceptbased affective vector space which we term, EmoSenticSpace. We then use this for deep learning-based sentiment analysis, in combination with our novel linguistic pattern-based affective reasoning method termed sentiment flow. Finally, we integrate all our text-based techniques and combine them with a novel deep learning-based visual feature extractor for multimodal sentiment analysis and emotion recognition. Comparative experimental results using a range of benchmark datasets have demonstrated the effectiveness of the proposed approach.

[pt] Os modelos ARCH e GARCH vêm sendo bastante explorados tanto tecnicamente quanto em estudos empíricos desde suas respectivas criações em 1982 e 1986. Contudo, o enfoque sempre foi na reprodução dos fatos estilizados das séries financeiras e na previsão de volatilidade, onde o GARCH(1,1) é o mais utilizado. Estudos sobre identificação dos modelos GARCH são muito raros. Diante desse contexto, este trabalho propõe um sistema inteligente para melhorar a identificação da correta especificação dos modelos GARCH, evitando assim o uso indiscriminado dos modelos GARCH(1,1). Para validar a eficácia do sistema proposto, séries simuladas foram utilizadas. Os resultados derivados desse sistema são comparados com os modelos escolhidos pelos critérios de informação AIC e BIC. O desempenho das previsões dos modelos identificados por esses métodos são comparados utilizando-se séries reais. / [en] ARCH and GARCH models have been largely explored technically and empirically since their creation in 1982 and 1986, respectively. However, the focus has always been on stylized facts of financial time series or volatility forecasts, where GARCH(1,1) has commonly been used. Studies on identification of GARCH models have been rare. In this context, this work aims to develop an intelligent system for improving the specification of GARCH models, thus avoiding the indiscriminate use of the GARCH(1,1) model. In order to validate the efficacy of the proposed system, simulated time series are used. Results are compared to chosen models through AIC and BIC criteria. Their performances are then compared by using real data.

Cidades Inteligentes são sistemas urbanos que usam as Tecnologias da Informação e Comunicação (TICs) para tornar a infraestrutura e os serviços públicos de uma cidade mais interativos, acessíveis e eficientes aos seus habitantes. Com isso, surge a necessidade de proporcionar novos tipos de serviços que busquem auxiliar na organização da cidade, no bem-estar das pessoas e auxiliem a melhorar a governança da cidade. Nesse contexto, o Sensoriamento Urbano é um paradigma emergente, situado no escopo de Cidades Inteligentes, que combina a ubiquidade de smartphones e de diferentes tipos de sensores para coletar dados que retratam diferentes aspectos da cidade. Um aspecto importante nesse processo de sensoriamento é a transmissão dos dados coletados para serem processados por um sistema centralizado remoto. Isso demanda uma ampla e constante cobertura de infraestrutura de rede de comunicação, fato que nem sempre é possível. As Redes Oportunistas surgem como uma alternativa complementar e inovadora para situações como essa, onde as aplicações precisam transmitir dados porém a infraestrutura de rede é intermitente ou mesmo inexistente. Entretanto, com o uso de Redes Oportunistas, as aplicações além de herdarem os seus benefícios também herdam os desafios existentes na área relacionados à tomada de decisão de encaminhamento das mensagens. Dessa forma, a presente tese busca responder ao questionamento de como intensificar a disseminação de conteúdo e o encaminhamento de mensagens em aplicações de Sensoriamento Urbano que fazem uso de Redes Oportunistas como paradigma complementar de comunicação. Para isso é proposto o Situs, um componente de software baseado em Redes Oportunistas que utiliza Sensibilidade à Situação com vistas à proatividade nas tomadas de decisões de roteamento para com isso intensificar a entrega de mensagens. Ele emprega Lógica Fuzzy para a compreensão da situação e uma rede neural chamada Echo State Network (ESN) para efetuar a projeção de situações. Os resultados experimentais demonstraram que a sua performance supera algumas das principais iniciativas existentes na literatura. Por fim, pode-se concluir que ele é capaz de preencher as lacunas do estado da arte apresentadas durante o desenvolvimento da tese sendo capaz de proporcionar um comportamento proativo com o uso de Sensibilidade à Situação. / Smart cities are urban systems that uses Information and Communication Technologies (ICTs) to make infrastructure and public services in a more interactive, accessible and efficient city to its inhabitants. With this comes the need to provide new types of services that seeks to assist in the organization of the city, the well-being of people and assist to improve the governance of the city. In this context, urban sensing is an emerging paradigm, sited in the Smart Cities scope, combining the ubiquity of smartphones with the capability of measuring o sensors to collect data that depict different aspects of the city. This ecosystem consists of different types of mobile and fixed devices orchestrated by a computational architecture that encompass the full sensing process. An important aspect of this process is the transmission of data collected for processing by a remote central system. This requires a broad and constant coverage of communication network infrastructure, a fact that is not always possible. The Opportunistic Networks emerge as an innovative and complementary alternative for situations like this where the applications needs to transmit data but the network infrastructure is intermittent or unavailable. However, using Opportunistic Networks, applications as well as inherit its benefits also inherit the existing challenges in the area related to decision-making of messages forwarding. Thus, this thesis seeks to answer the question of how to improve the dissemination of content and message routing of urban sensing applications that makes use of Opportunistic networks as complementary communication paradigm. Therefore, it proposes Situs, a software component based on Opportunistic Networks that uses Situation Awareness towards a proactivity in making routing decisions. For such task, it applies fuzzy logic for situation comprehension and a king of neural network called Echo State Network (ESN) for situation projection. The results of the experiments showed that their performance outperforms some existent initiatives in literature. Finally, we argue it fullfills the gaps of state of art presented in this thesis and could provide a proactive behaviour with usage of situation awareness.

Cidades Inteligentes são sistemas urbanos que usam as Tecnologias da Informação e Comunicação (TICs) para tornar a infraestrutura e os serviços públicos de uma cidade mais interativos, acessíveis e eficientes aos seus habitantes. Com isso, surge a necessidade de proporcionar novos tipos de serviços que busquem auxiliar na organização da cidade, no bem-estar das pessoas e auxiliem a melhorar a governança da cidade. Nesse contexto, o Sensoriamento Urbano é um paradigma emergente, situado no escopo de Cidades Inteligentes, que combina a ubiquidade de smartphones e de diferentes tipos de sensores para coletar dados que retratam diferentes aspectos da cidade. Um aspecto importante nesse processo de sensoriamento é a transmissão dos dados coletados para serem processados por um sistema centralizado remoto. Isso demanda uma ampla e constante cobertura de infraestrutura de rede de comunicação, fato que nem sempre é possível. As Redes Oportunistas surgem como uma alternativa complementar e inovadora para situações como essa, onde as aplicações precisam transmitir dados porém a infraestrutura de rede é intermitente ou mesmo inexistente. Entretanto, com o uso de Redes Oportunistas, as aplicações além de herdarem os seus benefícios também herdam os desafios existentes na área relacionados à tomada de decisão de encaminhamento das mensagens. Dessa forma, a presente tese busca responder ao questionamento de como intensificar a disseminação de conteúdo e o encaminhamento de mensagens em aplicações de Sensoriamento Urbano que fazem uso de Redes Oportunistas como paradigma complementar de comunicação. Para isso é proposto o Situs, um componente de software baseado em Redes Oportunistas que utiliza Sensibilidade à Situação com vistas à proatividade nas tomadas de decisões de roteamento para com isso intensificar a entrega de mensagens. Ele emprega Lógica Fuzzy para a compreensão da situação e uma rede neural chamada Echo State Network (ESN) para efetuar a projeção de situações. Os resultados experimentais demonstraram que a sua performance supera algumas das principais iniciativas existentes na literatura. Por fim, pode-se concluir que ele é capaz de preencher as lacunas do estado da arte apresentadas durante o desenvolvimento da tese sendo capaz de proporcionar um comportamento proativo com o uso de Sensibilidade à Situação. / Smart cities are urban systems that uses Information and Communication Technologies (ICTs) to make infrastructure and public services in a more interactive, accessible and efficient city to its inhabitants. With this comes the need to provide new types of services that seeks to assist in the organization of the city, the well-being of people and assist to improve the governance of the city. In this context, urban sensing is an emerging paradigm, sited in the Smart Cities scope, combining the ubiquity of smartphones with the capability of measuring o sensors to collect data that depict different aspects of the city. This ecosystem consists of different types of mobile and fixed devices orchestrated by a computational architecture that encompass the full sensing process. An important aspect of this process is the transmission of data collected for processing by a remote central system. This requires a broad and constant coverage of communication network infrastructure, a fact that is not always possible. The Opportunistic Networks emerge as an innovative and complementary alternative for situations like this where the applications needs to transmit data but the network infrastructure is intermittent or unavailable. However, using Opportunistic Networks, applications as well as inherit its benefits also inherit the existing challenges in the area related to decision-making of messages forwarding. Thus, this thesis seeks to answer the question of how to improve the dissemination of content and message routing of urban sensing applications that makes use of Opportunistic networks as complementary communication paradigm. Therefore, it proposes Situs, a software component based on Opportunistic Networks that uses Situation Awareness towards a proactivity in making routing decisions. For such task, it applies fuzzy logic for situation comprehension and a king of neural network called Echo State Network (ESN) for situation projection. The results of the experiments showed that their performance outperforms some existent initiatives in literature. Finally, we argue it fullfills the gaps of state of art presented in this thesis and could provide a proactive behaviour with usage of situation awareness.

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / The present work is a study of a methodology that seeks to be more economical and efficient compared to traditional methods, the design of three phase distribution transformers. To this end, we use a tool of Computational Intelligence, based on previous calculations made in the traditional way by manufacturing industries and books that address this matter. It is evident here, another way to design the active part of the transformer, the dimension of the core and coil, for this, used Genetic Algorithms (GA) which will assist in the calculation and selection of parameters, providing the best choice of design of the transformer, minimizing, on average, 5% in the cost and losses in 2.5% as input parameters the following standard. With changes in the fitness function of GA, one can simply adapt to new requirements and standards. As a result of using the genetic algorithm can find a design solution that is more efficient than that used by the manufacturers. The losses calculated values meet established by NBR-5440. / O presente trabalho faz um estudo de uma metodologia que busca ser mais econÃmica e eficiente, em comparaÃÃo com mÃtodos tradicionais, no projeto de transformadores de distribuiÃÃo trifÃsico. Para tal, utiliza-se uma ferramenta de InteligÃncia Computacional, tendo como base cÃlculos anteriores feitos de forma tradicional por indÃstrias fabricantes e livros que tratam desse assunto. Evidencia-se, aqui, mais uma maneira de projetar a parte ativa do transformador, ou seja, o dimensionamento do nÃcleo e bobinas; para isso, utiliza-se Algoritmos GenÃticos (GA), que irÃo auxiliar no cÃlculo e escolha dos parÃmetros, proporcionando a melhor escolha do projeto do transformador, minimizando, em mÃdia, o custo em 5 % e as perdas em 2,5 %, conforme parÃmetros de entrada seguindo a norma. Com modificaÃÃes na funÃÃo de fitness do GA, pode-se adequar de maneira simples a novas exigÃncias e normas. Como consequÃncia da utilizaÃÃo do Algoritmo GenÃtico pode-se encontrar uma soluÃÃo de projeto que à mais eficiente do que a utilizada pelos fabricantes. As perdas calculadas atendem valores estabelecidos pela norma NBR-5440.