No description available.

This thesis explores the problems raised by the aggregation of entities into a global, collective level, an old problem encountered in many fields of science. We work on three projects related to the aggregation problem in social systems, using tools derived from statistical physics, and more generally quantitative tools. The first project focus on a paradigmatic model of the emergence of puzzling macroscopic behavior from simple individual rules, Schelling's segregation model. We hence propose an analytical resolution of this model and we studied analytically and via simulations the effect of several forms of cooperation between individual agents on the collective behavior. These questions are tackled in a mutually beneficial way for both economics and physics. The second project is based on the exploration of huge databases on scientific literature. We hence produce several 'science maps' representing the fields of complex systems (its internal structure and coherence being analysed through the references used by ~140000 relevant articles) and the research carried out in a scientific institution such as the ENS de Lyon. Finally, the third project deals with the elaboration of models of social phenomena based on natural sciences tools but sociologically grounded. We hence present the elaboration process of a model built with a team of sociologists. We then propose an opinion model specifically designed to explore a single question: the existence of lasting structure from non lasting entities.

Over the past ten years, the architecture community has witnessed the end of single-threaded performance scaling and a subsequent shift in focus toward multicore and manycore processing. While this is an exciting time for architects, with many new opportunities and design spaces to explore, this brings with it some new challenges. One area that is especially impacted is the memory subsystem. Specifically, the design, verification, and evaluation of cache coherence protocols becomes very challenging as cores become more numerous and more diverse. This dissertation examines these issues and presents Manager-Client Pairing as a solution to the challenges facing next-generation coherence protocol design. By defining a standardized coherence communication interface and permissions checking algorithm, Manager-Client Pairing enables coherence hierarchies to be constructed and evaluated quickly without the high design-cost previously associated with hierarchical composition. Further, Manager-Client Pairing also allows for verification composition, even in the presence of protocol heterogeneity. As a result, this rapid development of diverse protocols is ensured to be bug-free, enabling architects to focus on performance optimization, rather than debugging and correctness concerns, while comparing diverse coherence configurations for use in future heterogeneous systems.

Submitted by Priscilla Araujo (priscilla@ibict.br) on 2018-10-03T17:49:45Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Cleiton da Mota de Souza_Mestrado_2018.pdf: 3867594 bytes, checksum: 0f665acec6936399da410c4a79ad9e8d (MD5) / Made available in DSpace on 2018-10-03T17:49:45Z (GMT). No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Cleiton da Mota de Souza_Mestrado_2018.pdf: 3867594 bytes, checksum: 0f665acec6936399da410c4a79ad9e8d (MD5) Previous issue date: 2018-03-19 / A Amazônia com sua vasta biota se caracteriza como um dos territórios mais ricos e menos explorados do planeta. A relevância deste bioma se torna ponto norteador para a questão acerca da responsabilidade assim como sobre as temáticas do conhecimento científico produzido sobre ele. Diante disto, o presente trabalho visa mapear os autores e temáticas das produções científicas sobre este bioma, objetivando responder à questão de pesquisa: o conhecimento científico sobre a Amazônia reflete as políticas públicas voltadas para a região? O estudo empírico, que é de caráter exploratório e quantitativo, embasado em técnicas bibliométricas, teve como fonte de informação primária os dados referenciais de artigos científicos. Estes dados foram coletados na base de dados Scopus, a partir do termo Amazon*, considerando o período de 1991 a 2015. Para a análise, os dados foram agrupados em cinco quinquênios: 1- 1991 a 1995, 2- 1996 a 2000, 3- 2001 a 2005, 4- 2006 a 2010 e 5- 2011 a 2015. Foram realizadas análises de coocorrências de palavra-chave visando à identificação de redes temáticas e também análise de coautoria para identificar as redes de colaboração entre os atores (países e autores). Para a elaboração dos mapas com as redes de autores, países e temático, foi utilizado o software VOSviewer, Como conclusões, observou-se um aumento gradativo das publicações sobre Amazônia nos períodos estudados. Ademais, as informações obtidas denotam um relativo distanciamento entre as políticas públicas voltadas para a região e o conhecimento científico produzido. Espera-se que o conjunto de resultados possa contribuir para a melhor compreensão das pesquisas realizadas e divulgadas sobre e oriundas da região. / The Amazon with its vast biota is characterized as one of the richest and least exploited territories on the planet. The relevance of this biome becomes the guiding point for the question about responsibility as well as about the themes of the scientific knowledge produced on it. In view of this, the present work aims to map the authors and themes of the scientific productions about this biome, aiming to answer the question of research: does the scientific knowledge about the Amazon reflect the public policies focused on the region? The empirical study, which is an exploratory and quantitative one, based on bibliometric techniques, had as primary information source the reference data of scientific articles. These data were collected in the Scopus database, from the term Amazon *, considering the period from 1991 to 2015. For the analysis, the data were grouped in five quinquennials: 1- 1991 to 1995, 2- 1996 to 2000, 3 - 2001 to 2005, 4- 2006 to 2010 and 5- 2011 to 2015. Keyword co-occurrences analyzes were performed aiming at the identification of thematic networks and also analysis of co-authorship to identify the networks of collaboration between the actors (countries and authors). For the elaboration of the maps with the networks of authors, countries and thematic, the software VOSviewer was used. As conclusions, there was a gradual increase of the publications about Amazonia in the studied periods. In addition, the information obtained shows a relative distance between the public policies focused on the region and the scientific knowledge produced. It is hoped that the results set will contribute to a better understanding of the research carried out and disseminated on and coming from the region.

Submitted by Rachel Pereira (rachelprr@yahoo.com.br) on 2015-12-16T16:54:28Z No. of bitstreams: 1 RobertoLovon_Tese-IBICT-Doutorado_2015.pdf: 4719896 bytes, checksum: 5aabc90c78a27d78fcda1b718f8ef32d (MD5) / Made available in DSpace on 2015-12-16T16:54:28Z (GMT). No. of bitstreams: 1 RobertoLovon_Tese-IBICT-Doutorado_2015.pdf: 4719896 bytes, checksum: 5aabc90c78a27d78fcda1b718f8ef32d (MD5) Previous issue date: 2015-08-26 / No Brasil, a maior parte da atividade científica é desenvolvida nos programas de pós-graduação das universidades públicas. A diversidade temática e a complexidade estrutural dessas instituições fazem com que a identificação de seus perfis e padrões acadêmicos não seja uma tarefa fácil. O estudo da produção científica, a partir da análise de domínios científicos, permite conhecer a forma como estão configuradas/estruturadas as atividades científicas dessas instituições e os diversos atores nelas envolvidos. Considerando as seguintes questões de pesquisa: De que maneira estão configurados os domínios científicos da UFRJ? Eles são estáticos ou mudam ao longo do tempo? Os diversos domínios interagem entre si? Esta tese tem como principal objetivo mapear os domínios científicos de uma das maiores e mais importantes instituições de ensino e pesquisa do país, a Universidade Federal do Rio de Janeiro (UFRJ), identificando as principais áreas de conhecimento da instituição, bem como as interações entre os diferentes atores de seu corpo social. Trata-se de um estudo de natureza quantitativa, envolvendo informações de áreas vinculadas à produção científica e autores dessa produção dos programas de pós-graduação da UFRJ, que utiliza técnicas bibliométricas e de análise de redes a partir das medidas de co-ocorrência. Para tanto, este trabalho investigou 44.233 registros recuperados da base de dados institucional, EspaçoSIGMA.UFRJ, de artigos publicados em periódicos no período de 2001 a 2012. Os dados foram organizados e analisados com auxílio dos softwares Excel e Gephi. A partir de dois blocos de análises, considerando artigos em periódicos-áreas do conhecimento e artigos em periódicos-autores, os resultados mostram as principais áreas temáticas da UFRJ, destacando-se com maior quantidade de registros: Medicina, Química, Bioquímica, Zoologia, Microbiologia, Engenharia Química, Engenharia de Materiais e Metalúrgica. Além destas, encontramos também destacada presença de registros nas áreas de Letras, História, Educação, Psicologia. Outro aspecto observado foi o nível de interação entre as áreas: em torno de 30% dos artigos publicados são vinculados a mais de uma área de conhecimento. Verificou-se uma dinâmica de interação entre áreas de conhecimento, na qual se destacam: Ciências da Saúde, Engenharias e Ciências Humanas. O desdobramento das grandes áreas permitiu identificar aspectos mais específicos da interação entre as mesmas, possibilitando o mapeamento de comunidades de áreas de conhecimento. A análise por autores e as interações que se estabelecem a partir da medida de coautoria revelaram uma tendência de queda no número de artigos contendo entre 1 e 4 autores, enquanto que tende a crescer o número de artigos contendo 5 ou mais autores. A distribuição das autorias pelo tipo de vínculo institucional do autor e a relação entre produtividade, colaboração (coautoria) e ligação institucional também foram observados. As métricas de análise de redes sociais mostraram que, ao longo do período estudado, novas colaborações entre autores da UFRJ foram estabelecidas, formando novos agrupamentos/comunidades de autores. As maiores comunidades de autores identificadas foram detalhadas, revelando suas principais especificações. / In Brazil, most of the scientific activity is developed by graduate programs from public universities. The thematic diversity and structural complexity of these institutions make the identification of their profiles and academic standards a hard task. The study of scientific production, using scientific domain analysis, allows knowing how the scientific activities are configured and the various actors involved. Considering the research questions: how are the scientific domains of the UFRJ configured? Do they change over time? Do different domains interact with each other? This thesis aims mapping the scientific domains of one of the largest and most important educational and research institutions of the country, the Federal University of Rio de Janeiro (UFRJ), identifying the its main fields of expertise as well as the interactions between the different actors of its social body. It is a quantitative study that encompasses information related to the fields of knowledge and authorship from the scientific production of UFRJ’s graduate programs, which uses the network analysis technique based on the co-occurrence measures. The study investigated 44,233 records retrieved from the institutional database, EspaçoSIGMA.UFRJ, articles published in journals from 2001 to 2012. Data were organized and analyzed using Excel and Gephi software. The two sets of analysis, including articles in journals-fields of knowledge and articles in journal-authors, the results show the main thematic fields of UFRJ. Those with the highest number of records are: Medicine, Chemistry, Biochemistry, Zoology, Microbiology, Chemical Engineering, Materials Engineering and Metallurgy. We also find a strong presence of records in the fields of Letters, History, Education, Psychology. Another finding was the level of interaction among fields: around 30% of the published articles are linked to more than one field of knowledge. From this perspective, there was a dynamic interaction, over the period studied, among fields, in which the largest core fields are: Health Sciences, Engineering and Humanities. The analysis of large fields has identified more specific aspects of the interaction between them, allowing the mapping of clusters/communities of fields of knowledge. The analysis by authorship and the interactions established from co-authorship measure showed a downward trend in the number of articles with 1 to 4 authors, whereas tends to increase the number of articles with 5 or more authors. The distribution of authorship by the type of institutional link of the author and the relationship between productivity, collaboration (co-author) and institutional link were also observed. The social network analysis showed that, over the study period, new collaborations among authors of UFRJ were established, forming new clusters/communities of authors. The largest authors’ clusters/communities were detailed and revealed their main specifications. It is, therefore, a singular study, with a comprehensive view of the scientific activity of the entire social body of one of the leading teaching and research institutions in Brazil.

L'objectif principal de la thèse consiste à explorer des problématiques propres aux sciences sociales et à les étudier à l'aide d'outils issus du champ de la physique statistiques et des systèmes complexes. Le travail de la thèse s'est ainsi décliné sur trois grands sujets dont la problématique principale est la question de l'agrégation d'entités individuelles en une structure collective. Le premier sujet est centré sur un exemple paradigmatique de l'émergence d'un comportement collectif macroscopique inattendu à partir de règles individuelles simples: le modèle de ségrégation de Schelling. Nous avons notamment proposé une résolution analytique inédite de ce modèle et nous avons étudié analytiquement et via des simulations l'impact de différentes formes de coopération entre agents individuels sur le comportement collectif global. Cette thématique a été développée à la fois d'un point de vue économique et d'un point de vue physique. Le second sujet porte sur l'exploration de bases de données bibliométriques. Nous avons ainsi produit des 'cartes des sciences' représentant le champ des systèmes complexes (sa structure interne étant décrypté via une analyse des références utilisées par ~140000 articles) ou encore l'état de la recherche au sein d'un établissement tel que l'ENS de Lyon. Enfin, le troisième thème porte sur l'élaboration de modèles basés sur des outils des sciences 'dures' mais sociologiquement fondés. Nous présentons ainsi le processus d'élaboration d'un modèle construit avec une équipe de sociologues. Enfin, nous développons un modèle d'opinion répondant spécifiquement à une question: l'existence de structures qui durent à partir d'entités qui ne durent pas. / This thesis explores the problems raised by the aggregation of entities into a global, collective level, an old problem encountered in many fields of science. We work on three projects related to the aggregation problem in social systems, using tools derived from statistical physics, and more generally quantitative tools. The first project focus on a paradigmatic model of the emergence of puzzling macroscopic behavior from simple individual rules, Schelling's segregation model. We hence propose an analytical resolution of this model and we studied analytically and via simulations the effect of several forms of cooperation between individual agents on the collective behavior. These questions are tackled in a mutually beneficial way for both economics and physics. The second project is based on the exploration of huge databases on scientific literature. We hence produce several 'science maps' representing the fields of complex systems (its internal structure and coherence being analysed through the references used by ~140000 relevant articles) and the research carried out in a scientific institution such as the ENS de Lyon. Finally, the third project deals with the elaboration of models of social phenomena based on natural sciences tools but sociologically grounded. We hence present the elaboration process of a model built with a team of sociologists. We then propose an opinion model specifically designed to explore a single question: the existence of lasting structure from non lasting entities.