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

Sobre a natureza dos processos físicos de aquecimento da atmosfera de estrelas do tipo solar

In the present work, we have studied the nature of the physical processes of the coronal heating, considering as basis significant samples of single and binary evolved stars, that have been achieved with the ROSAT satellite. In a total of 191 simple stars were studied, classified in the literature as giants with spectral type F, G and K. The results were compared with those obtained from 106 evolved stars of spectral type F, G and K, which belong to the spectroscopic binary systems. Accurate measurements on rotation and information about binarity were obtained from De Medeiros s catalog. We have analysed the behavior of the coronal activity in function of diverse stellar parameters. With the purpose to better clarify the profile of the stars evolution, the HR diagram was built for the two samples of stars, the single and the binary ones. The evolved traces added in the diagram were obtained from the Toulouse-Geneve code, Nascimento et al. (2000). The stars were segregated in this diagram not only in range of rotational speed but also in range of X-ray flux. Our analysis shows clearly that the single stars and the binary ones have coronal activity controlled by physical process independent on the rotation. Non magnetic processes seem to be strongly influencing the coronal heating. For the binary stars, we have also studied the behavior of the coronal emission as a function of orbital parameters, such as period and eccentricity, in which it was revealed the existence of a discontinuity in the emission of X-rays around an orbital period of 100 days. The study helped to conclude that circular orbits of the binary stars are presented as a necessary property for the existence of a higher level ofX-rays emission, suggesting that the effect of the gravitational tide has an important role in the coronal activity level. When applied the Kolmogorov-Smirnov test (KS test ) for the Vsini and FX parameters to the samples of single and binary stars, we could evidence very relevant aspects for the understanding of the mechanisms inherent to the coronal activity. For the Vsini parameter, the differences between the single stars and the binary ones for rotation over 6.3 km/s were really remarkable. We believe, therefore, that the existence of gravitational tide is, at least, one of the factors that most contribute for this behavior. About the X-rays flux, the KS test showed that the behavior of the single and the binary stars, regarding the coronal activity, comes from the same origin

Deciding difference logic in a Nelson-Oppen combination framework

O método de combinação de Nelson-Oppen permite que vários procedimentos de decisão, cada um projetado para uma teoria específica, possam ser combinados para inferir sobre teorias mais abrangentes, através do princípio de propagação de igualdades. Provadores de teorema baseados neste modelo são beneficiados por sua característica modular e podem evoluir mais facilmente, incrementalmente. Difference logic é uma subteoria da aritmética linear. Ela é formada por constraints do tipo x − y ≤ c, onde x e y são variáveis e c é uma constante. Difference logic é muito comum em vários problemas, como circuitos digitais, agendamento, sistemas temporais, etc. e se apresenta predominante em vários outros casos. Difference logic ainda se caracteriza por ser modelada usando teoria dos grafos. Isto permite que vários algoritmos eficientes e conhecidos da teoria de grafos possam ser utilizados. Um procedimento de decisão para difference logic é capaz de induzir sobre milhares de constraints. Um procedimento de decisão para a teoria de difference logic tem como objetivo principal informar se um conjunto de constraints de difference logic é satisfatível (as variáveis podem assumir valores que tornam o conjunto consistente) ou não. Além disso, para funcionar em um modelo de combinação baseado em Nelson-Oppen, o procedimento de decisão precisa ter outras funcionalidades, como geração de igualdade de variáveis, prova de inconsistência, premissas, etc. Este trabalho apresenta um procedimento de decisão para a teoria de difference logic dentro de uma arquitetura baseada no método de combinação de Nelson-Oppen. O trabalho foi realizado integrando-se ao provador haRVey, de onde foi possível observar o seu funcionamento. Detalhes de implementação e testes experimentais são relatados

Analisando o desempenho do ClassAge: um sistema multiagentes para classificação de padrões

The use of multi-agent systems for classification tasks has been proposed in order to overcome some drawbacks of multi-classifier systems and, as a consequence, to improve performance of such systems. As a result, the NeurAge system was proposed. This system is composed by several neural agents which communicate and negotiate a common result for the testing patterns. In the NeurAge system, a negotiation method is very important to the overall performance of the system since the agents need to reach and agreement about a problem when there is a conflict among the agents. This thesis presents an extensive analysis of the NeurAge System where it is used all kind of classifiers. This systems is now named ClassAge System. It is aimed to analyze the reaction of this system to some modifications in its topology and configuration