Uma implementação paralela híbrida para o problema do caixeiro viajante usando algoritmos genéticos, GRASP e aprendizagem por reforço

The metaheuristics techiniques are known to solve optimization problems classified as NP-complete and are successful in obtaining good quality solutions. They use non-deterministic approaches to generate solutions that are close to the optimal, without the guarantee of finding the global optimum. Motivated by the difficulties in the resolution of these problems, this work proposes the development of parallel hybrid methods using the reinforcement learning, the metaheuristics GRASP and Genetic Algorithms. With the use of these techniques, we aim to contribute to improved efficiency in obtaining efficient solutions. In this case, instead of using the Q-learning algorithm by reinforcement learning, just as a technique for generating the initial solutions of metaheuristics, we use it in a cooperative and competitive approach with the Genetic Algorithm and GRASP, in an parallel implementation. In this context, was possible to verify that the implementations in this study showed satisfactory results, in both strategies, that is, in cooperation and competition between them and the cooperation and competition between groups. In some instances were found the global optimum, in others theses implementations reach close to it. In this sense was an analyze of the performance for this proposed approach was done and it shows a good performance on the requeriments that prove the efficiency and speedup (gain in speed with the parallel processing) of the implementations performed

Calibração cega de receptores cinco-portas baseada em separação cega de fontes

The exponential growth in the applications of radio frequency (RF) is accompanied by great challenges as more efficient use of spectrum as in the design of new architectures for multi-standard receivers or software defined radio (SDR) . The key challenge in designing architecture of the software defined radio is the implementation of a wide-band receiver, reconfigurable, low cost, low power consumption, higher level of integration and flexibility. As a new solution of SDR design, a direct demodulator architecture, based on fiveport technology, or multi-port demodulator, has been proposed. However, the use of the five-port as a direct-conversion receiver requires an I/Q calibration (or regeneration) procedure in order to generate the in-phase (I) and quadrature (Q) components of the transmitted baseband signal. In this work, we propose to evaluate the performance of a blind calibration technique without additional knowledge about training or pilot sequences of the transmitted signal based on independent component analysis for the regeneration of I/Q five-port downconversion, by exploiting the information on the statistical properties of the three output signals

ISE-SPL: Uma abordagem baseada em linha de produtos de software aplicada à geração automática de sistemas para educação médica na plataforma E-learning

E-learning, which refers to the use of Internet-related technologies to improve knowledge and learning, has emerged as a complementary form of education, bringing advantages such as increased accessibility to information, personalized learning, democratization of education and ease of update, distribution and standardization of the content. In this sense, this paper aims to develop a tool, named ISE-SPL, whose purpose is the automatic generation of E-learning systems for medical education, making use of concepts of Software Product Lines. It consists of an innovative methodology for medical education that aims to assist professors of healthcare in their teaching through the use of educational technologies, all based on computing applied to healthcare (Informatics in Health). The tests performed to validate the ISE-SPL were divided into two stages: the first was made by using a software analysis tool similar to ISE-SPL, called SPLOT and the second was performed through usability questionnaires to healthcare professors who used ISESPL. Both tests showed positive results, proving it to be an efficient tool for generation of E-learning software and useful for professors in healthcare

Classificação de padrões através de um comitê de máquinas aprimorado por aprendizagem por reforço

Reinforcement learning is a machine learning technique that, although finding a large number of applications, maybe is yet to reach its full potential. One of the inadequately tested possibilities is the use of reinforcement learning in combination with other methods for the solution of pattern classification problems. It is well documented in the literature the problems that support vector machine ensembles face in terms of generalization capacity. Algorithms such as Adaboost do not deal appropriately with the imbalances that arise in those situations. Several alternatives have been proposed, with varying degrees of success. This dissertation presents a new approach to building committees of support vector machines. The presented algorithm combines Adaboost algorithm with a layer of reinforcement learning to adjust committee parameters in order to avoid that imbalances on the committee components affect the generalization performance of the final hypothesis. Comparisons were made with ensembles using and not using the reinforcement learning layer, testing benchmark data sets widely known in area of pattern classification

Desenvolvimento de software para simulação de motores com dispositivos de partida baseada na integração do atp com o toprede

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior

Algoritmos genéticos aplicados a um comitê de LS-SVM em problemas de classificação

The pattern classification is one of the machine learning subareas that has the most outstanding. Among the various approaches to solve pattern classification problems, the Support Vector Machines (SVM) receive great emphasis, due to its ease of use and good generalization performance. The Least Squares formulation of SVM (LS-SVM) finds the solution by solving a set of linear equations instead of quadratic programming implemented in SVM. The LS-SVMs provide some free parameters that have to be correctly chosen to achieve satisfactory results in a given task. Despite the LS-SVMs having high performance, lots of tools have been developed to improve them, mainly the development of new classifying methods and the employment of ensembles, in other words, a combination of several classifiers. In this work, our proposal is to use an ensemble and a Genetic Algorithm (GA), search algorithm based on the evolution of species, to enhance the LSSVM classification. In the construction of this ensemble, we use a random selection of attributes of the original problem, which it splits the original problem into smaller ones where each classifier will act. So, we apply a genetic algorithm to find effective values of the LS-SVM parameters and also to find a weight vector, measuring the importance of each machine in the final classification. Finally, the final classification is obtained by a linear combination of the decision values of the LS-SVMs with the weight vector. We used several classification problems, taken as benchmarks to evaluate the performance of the algorithm and compared the results with other classifiers

Repensando os distúrbios de aprendizagem a partir da psicologia histórico-cultural

A psicologia histórico-cultural assume que o fator biológico determina a base das reações inatas dos indivíduos. Sobre esta base se constitui todo o sistema de reações adquiridas, sendo estas determinadas mais pela estrutura do meio cultural da criança do que pelas disposições biológicas. Se é por meio do processo de apropriação da cultura que cada homem adquire as capacidades humanas, a compreensão atual acerca dos distúrbios de aprendizagem pode ser reconfigurada, demonstrando que mediações adequadas e consistentes podem ter caráter revolucionário para a aprendizagem, ao tornarem presente o talento cultural quando o talento biológico não se revela como esperado.

A Informática como recurso didático para a aprendizagem de Física no ensino médio

The accomplishment of this work was motivated by our concerns, while teacher of Physics discipline, in the medium teaching and in the higher education, where we have been observing a lot of difficulties in the students' acting, to the they try to give pursuit out to their studies of the classroom, because of the lack of appropriate equipments, or even, of a laboratory where they can put in practice the studied contents. The work aims at to build and to test an educational software that it serves as tool auxiliadora in Physics learning in the Medium Teaching, in the area of electricity, with emphasis in the study of the electrodynamic in resistors. The developed software comes as an alternative to the learning problems, putting the computer science as auxiliary tool, because, besides being an alternative in the middle of the technological expansion, endowed with several resources, it stimulates the significant learning, according to David Ausubel's perspective. A software containing a program destined to the applicability of physics contents in the branch of the electricity is presented as an auxiliary tool, where the student can, not just, to review the contents presented at room, as well as for in practice, through a virtual laboratory, some of these contents, besides testing their knowledge through a bank of discursive subjects. The evaluation of the developed software was made submitting him/it to the professionals' of physics area appreciation. Equally, through continuous evaluations, they were made comparisons among the students' of five different groups acting, in the same school, that you/they were used of the program as tool of his auxiliary learning, and the acting of those that didn't use it

Uma proposta de ensino e aprendizagem de Língua Inglesa integrada ao uso das TICS para o desenvolvimento da autonomia dos aprendizes

This study focuses on the teaching and learning of English based on a new theoretical perspective concerning the Zone of Proximal Development (ZPD), which points to the apprehension of metacognitive processes as the way to reach learning autonomy. The theoretical set underline this study is those of a new pedagogy based on the symbiosis of Bruner‟s (2002) and Freire‟s (2009) concepts and in the metacognitive theory. In a social context dominated by the Communication and Information Technologies (CIT), the integration between teacher and learners with the support of the internet has been used as the way to operationalize this new emergent proposal of the theoretical perspectives. The analyses have been conducted through Systemic and Integral Action Research. The results at the end of this study corroborate our hypothesis that the enlargement of spontaneous knowledge of the learners can facilitate the understanding of scientific concepts, stimulating their metacognition and thus promoting their autonomy.

Investigação da influência do cochilo na aprendizagem escolar

There is a known positive effect of nocturnal sleep for brain plasticity and the consolidation of declarative and procedural memories, as well as for the facilitation of insight in problem solving. However, a possible pro-mnemonic effect of daytime naps after learning is yet to be properly characterized. The goal of this project was to evaluate the influence of daytime naps on learning among elementary and middleschool students, measuring the one-day (acute), and semester-long (chronic) effects of post-learning naps on performance. In the Acute Day-Nap condition, the elementary students were exposed to a class and then randomly divided into three groups: Nap (N), Game-based English Class (GBEC) and Traditional English Class (TEC). There were 2 multiple-choice follow-up tests to evaluate students performance in the short and long runs. In the short run, the N group outperformed the other two groups; and such tendency was maintained in the long run. In the chronic condition, the middle-school students were randomly separated into two groups: Nap (N) and Class (C) and were observed during one academic term. The N group had increased school performance in relation to the C group. In the statistical analyses, independent t-tests were applied considering the scores significant when p<0,05, expressed in terms of average ± average standard error. Results can be interpreted as an indication that a single daytime nap opportunity is not enough to ensure learning benefits. Therefore, more research is needed in order to advocate in favor of a daytime nap as a pedagogical means of promoting enhanced school performance

Estilos de aprendizagem sob a ótica da Psicologia Evolucionista

Among a variety of learning conceptions, David Kolb´s Experiential Learning Theory proposes four different learning styles: diverging, characterized by orientation towards people and multi-perspective vision; assimilating, concerned with ideas and abstract concepts as well as theory formulation; converging, expert in dealing with technical tasks and problem solving; and accommodating, risk taker and good at getting things done. Interesting correlations have been pointed out between Kolb s learning styles, professional careers and genders. With respect to behaviors, specific cognitive skills and interests, sex differences are widely known, and explained by Evolutionary Psychology as the result of distinct selective pressures acting on each gender. The aim of this research was to assess adolescents learning styles and their relation with interests on school and career choices, analyzing possible gender differences. We distributed questionnaires to 221 senior high school students to research their preferences for school disciplines, professional activities and career choices. The Learning Style Inventory specified the learning style of each individual. Our results showed a high frequency of reflective styles, with predominance of females as diverging and males as assimilating. Concerning school and professional interests, there were correlations between styles oriented towards the abstract and technical interests. Moreover, females preferred disciplines related to languages and interpersonal activities while males preferred disciplines related to science and technical activities. There were more males in exact science and engineering careers, and more females in social science and applied social science. Correlations found between learning styles, school and professional interests corroborate Kolb´s propositions, and the findings about gender differences are supported by Evolutionary Psychology theories

Fluxo de carga continuado utilizando estratégia de parametrização geométrica baseada em parâmetros físicos

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Aprendizagem na interação ensino-serviço-comunidade: a formação na perspesctiva dialógica com a sociedade

Considerando o papel do ensino superior em saúde para a sociedade brasileira, em que os egressos dos cursos ofertados na área serão os profissionais prestadores de serviços à população, este estudo objetivou analisar o componente ensino do PET-Saúde da Família - Natal-RN na formação de estudantes dos cursos de graduação em saúde da UFRN. Foi realizada pesquisa qualitativa com análise de documentos das disciplinas SACI e POTI datados entre 2009 e 2011 (portfólios, avaliações de desempenho e oficina de avaliação), mediante o auxílio do software Alceste© e análise de conteúdo, segundo Bardin. Na análise foi encontrado como potencialidades: o alunato trabalhando em grupo tutorial multidisciplinar, cuja interação e contato com a Unidade de Saúde da Família, incluídos os profissionais, bem como a comunidade, instiga nos aprendizes o diálogo consigo mesmo e com o outro, numa construção dos ser/agir no mundo. Os textos trabalhados durante as aulas permitem refletir e teorizar a respeito da realidade observada, auxiliando-os na identificação dos problemas e no traçar estratégias de intervenção. Já a observação da realidade reveste o aluno de humanização. Este passa a captar as necessidades e dificuldades enfrentadas pela comunidade observada. Nas fragilidades ficaram evidenciados: problemas de relações interpessoais entre os estudantes da SACI; a maioria dos projetos de intervenção ocorrendo numa perspectiva paternalista, reproduzindo o modelo de prestação de serviços na saúde mais praticado nas sociedades brasileiras; dificuldades em aprofundar no aprendiz, a importância da teorização dos assuntos; problemas de financiamento de projetos de intervenção; descumprimento do plano de ensino em alguns grupos tutoriais; e, por fim, dificuldades dos alunos e monitores em acompanhar as atividade de pesquisa e extensão do PET-Saúde, pela falta de integração dos projetos pedagógicos dos cursos. Conclui-se que o componente ensino do PET-Saúde da Família adota metodologias ativas de ensino na inserção de alunos na Atenção Primária em Saúde, proporcionando uma formação dentro de princípios éticos e humanísticos a partir do trabalho em equipe e da inclusão reflexiva dos alunos na Estratégia Saúde da Família. Apesar da existência de fragilidades concernentes às relações interpessoais, descompasso entre as proposições multiprofissionais e interdisciplinares da SACI e POTI e as dificuldades de pô-las em prática em currículos fragmentados e organizados por disciplinas pouco flexíveis, potencialmente, ao fim dessas experiências conectadas a Atenção Primária, os discentes apresentam uma nova visão do cuidado com a saúde, próxima às necessidades da população, iniciando uma tomada de postura crítica e reflexiva, entendendo-se com sujeitos ativos no construir a saúde coletivamente

Otimização em comitês de classificadores: uma abordagem baseada em filtro para seleção de subconjuntos de atributos

Traditional applications of feature selection in areas such as data mining, machine learning and pattern recognition aim to improve the accuracy and to reduce the computational cost of the model. It is done through the removal of redundant, irrelevant or noisy data, finding a representative subset of data that reduces its dimensionality without loss of performance. With the development of research in ensemble of classifiers and the verification that this type of model has better performance than the individual models, if the base classifiers are diverse, comes a new field of application to the research of feature selection. In this new field, it is desired to find diverse subsets of features for the construction of base classifiers for the ensemble systems. This work proposes an approach that maximizes the diversity of the ensembles by selecting subsets of features using a model independent of the learning algorithm and with low computational cost. This is done using bio-inspired metaheuristics with evaluation filter-based criteria

Avaliação e intervenção da disortografia baseada na semiologia dos erros: revisão da literatura

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)