Algoritmo Q-learning como estratégia de exploração e/ou explotação para metaheurísticas GRASP e algoritmo genético

Techniques of optimization known as metaheuristics have achieved success in the resolution of many problems classified as NP-Hard. These methods use non deterministic approaches that reach very good solutions which, however, don t guarantee the determination of the global optimum. Beyond the inherent difficulties related to the complexity that characterizes the optimization problems, the metaheuristics still face the dilemma of xploration/exploitation, which consists of choosing between a greedy search and a wider exploration of the solution space. A way to guide such algorithms during the searching of better solutions is supplying them with more knowledge of the problem through the use of a intelligent agent, able to recognize promising regions and also identify when they should diversify the direction of the search. This way, this work proposes the use of Reinforcement Learning technique - Q-learning Algorithm - as exploration/exploitation strategy for the metaheuristics GRASP (Greedy Randomized Adaptive Search Procedure) and Genetic Algorithm. The GRASP metaheuristic uses Q-learning instead of the traditional greedy-random algorithm in the construction phase. This replacement has the purpose of improving the quality of the initial solutions that are used in the local search phase of the GRASP, and also provides for the metaheuristic an adaptive memory mechanism that allows the reuse of good previous decisions and also avoids the repetition of bad decisions. In the Genetic Algorithm, the Q-learning algorithm was used to generate an initial population of high fitness, and after a determined number of generations, where the rate of diversity of the population is less than a certain limit L, it also was applied to supply one of the parents to be used in the genetic crossover operator. Another significant change in the hybrid genetic algorithm is the proposal of a mutually interactive cooperation process between the genetic operators and the Q-learning algorithm. In this interactive/cooperative process, the Q-learning algorithm receives an additional update in the matrix of Q-values based on the current best solution of the Genetic Algorithm. The computational experiments presented in this thesis compares the results obtained with the implementation of traditional versions of GRASP metaheuristic and Genetic Algorithm, with those obtained using the proposed hybrid methods. Both algorithms had been applied successfully to the symmetrical Traveling Salesman Problem, which was modeled as a Markov decision process

Análise comparativa do desempenho de um Controlador Fuzzy acoplado a um PID Neural sintonizado por um Algoritmo Genético com Controladores Inteligentes Convencionais

On this paper, it is made a comparative analysis among a controller fuzzy coupled to a PID neural adjusted by an AGwith several traditional control techniques, all of them applied in a system of tanks (I model of 2nd order non lineal). With the objective of making possible the techniques involved in the comparative analysis and to validate the control to be compared, simulations were accomplished of some control techniques (conventional PID adjusted by GA, Neural PID (PIDN) adjusted by GA, Fuzzy PI, two Fuzzy attached to a PID Neural adjusted by GA and Fuzzy MISO (3 inputs) attached to a PIDN adjusted by GA) to have some comparative effects with the considered controller. After doing, all the tests, some control structures were elected from all the tested techniques on the simulating stage (conventional PID adjusted by GA, Fuzzy PI, two Fuzzy attached to a PIDN adjusted by GA and Fuzzy MISO (3 inputs) attached to a PIDN adjusted by GA), to be implemented at the real system of tanks. These two kinds of operation, both the simulated and the real, were very important to achieve a solid basement in order to establish the comparisons and the possible validations show by the results

Separação cega de fontes lineares e não lineares usando algoritmo genético, redes neurais artificiais RBF e negentropia de Rényi como medida de independência

Conventional methods to solve the problem of blind source separation nonlinear, in general, using series of restrictions to obtain the solution, often leading to an imperfect separation of the original sources and high computational cost. In this paper, we propose an alternative measure of independence based on information theory and uses the tools of artificial intelligence to solve problems of blind source separation linear and nonlinear later. In the linear model applies genetic algorithms and Rényi of negentropy as a measure of independence to find a separation matrix from linear mixtures of signals using linear form of waves, audio and images. A comparison with two types of algorithms for Independent Component Analysis widespread in the literature. Subsequently, we use the same measure of independence, as the cost function in the genetic algorithm to recover source signals were mixed by nonlinear functions from an artificial neural network of radial base type. Genetic algorithms are powerful tools for global search, and therefore well suited for use in problems of blind source separation. Tests and analysis are through computer simulations

Estudo avaliativo de um algoritmo genético auto-organizável e multiobjetivo utilizando aprendizado de máquina para aplicações de telecomunicações

This paper presents an evaluative study about the effects of using a machine learning technique on the main features of a self-organizing and multiobjective genetic algorithm (GA). A typical GA can be seen as a search technique which is usually applied in problems involving no polynomial complexity. Originally, these algorithms were designed to create methods that seek acceptable solutions to problems where the global optimum is inaccessible or difficult to obtain. At first, the GAs considered only one evaluation function and a single objective optimization. Today, however, implementations that consider several optimization objectives simultaneously (multiobjective algorithms) are common, besides allowing the change of many components of the algorithm dynamically (self-organizing algorithms). At the same time, they are also common combinations of GAs with machine learning techniques to improve some of its characteristics of performance and use. In this work, a GA with a machine learning technique was analyzed and applied in a antenna design. We used a variant of bicubic interpolation technique, called 2D Spline, as machine learning technique to estimate the behavior of a dynamic fitness function, based on the knowledge obtained from a set of laboratory experiments. This fitness function is also called evaluation function and, it is responsible for determining the fitness degree of a candidate solution (individual), in relation to others in the same population. The algorithm can be applied in many areas, including in the field of telecommunications, as projects of antennas and frequency selective surfaces. In this particular work, the presented algorithm was developed to optimize the design of a microstrip antenna, usually used in wireless communication systems for application in Ultra-Wideband (UWB). The algorithm allowed the optimization of two variables of geometry antenna - the length (Ls) and width (Ws) a slit in the ground plane with respect to three objectives: radiated signal bandwidth, return loss and central frequency deviation. These two dimensions (Ws and Ls) are used as variables in three different interpolation functions, one Spline for each optimization objective, to compose a multiobjective and aggregate fitness function. The final result proposed by the algorithm was compared with the simulation program result and the measured result of a physical prototype of the antenna built in the laboratory. In the present study, the algorithm was analyzed with respect to their success degree in relation to four important characteristics of a self-organizing multiobjective GA: performance, flexibility, scalability and accuracy. At the end of the study, it was observed a time increase in algorithm execution in comparison to a common GA, due to the time required for the machine learning process. On the plus side, we notice a sensitive gain with respect to flexibility and accuracy of results, and a prosperous path that indicates directions to the algorithm to allow the optimization problems with "η" variables

Melhoramento genético da mamona visando incremento de flores femininas

O objetivo deste trabalho foi testar métodos de seleção visando ao aumento de flores femininas na população FCA-UNESP-PB de mamona (Ricinus communis L.). A seleção foi realizada no município de Botucatu (SP), na safrinha de 2007. Por meio de seleção massal, foram selecionadas plantas com racemo primário estritamente feminino. Destas plantas, as que tinham reversão sexual foram autofecundadas. As avaliações foram realizadas na safrinha de 2008 em Botucatu e São Manuel (SP), onde foram comparados os tratamentos: método de seleção massal; método de seleção massal com autofecundação e testemunha (racemos de plantas colhidos ao acaso, sem seleção). Foram avaliados: porcentagem de flores femininas do racemo primário (%), produtividade de grãos (kg ha-1) e teor de óleo das sementes (%). O delineamento experimental utilizado foi o de blocos casualizados com 30 repetições. Os dados foram submetidos à análise de variância individual para cada local e conjuntamente para os dois locais, pelo teste F a 1% de probabilidade. Mediante os resultados conclui- se que o método de seleção massal com autofecundação foi aquele que proporcionou maiores valores de porcentagem de flores femininas no racemo primário, com ganho fenotípico realizado de 18% em Botucatu e 29% em São Manuel (SP). Por meio dos métodos de seleção, notou-se comportamento diferencial em relação aos locais para a característica produtividade de grãos, e o método seleção massal com autofecundação proporcionou a menor produtividade. No teor de óleo não houve diferenças significativas entre os métodos e os locais avaliados.

Utilização de análise de segregantes agrupados na identificação de marcadores ligados a genes que controlam a resistência à ferrugem (Puccinia psidii Winter) em Eucalyptus sp.

Devido a grande importância da cultura de Eucalyptus no Brasil, empresas do setor florestal têm buscado através de programas de melhoramento genético, reduzir as perdas de produção e atender a demanda do mercado de papel e celulose. Um exemplo, é a busca por genes de resistência a doenças, principalmente a ferrugem causada por Puccinia psidii Winter, que resulta em redução da produtividade em plantas altamente suscetíveis. No presente trabalho, mudas de Eucalyptus pertencentes a uma geração F1, provenientes do cruzamento controlado entre parentais híbridos E. grandis X E. urophylla, sendo eles resistente e suscetível, foram inoculadas com Puccinia psidii em casa de vegetação e acompanhadas até o aparecimento dos sintomas da ferrugem. Foram classificadas, em dois grupos: resistentes (ausência de sintomas) e suscetíveis (presença de sintomas e esporulação). As amostras de DNA foram comparadas com o uso de marcadores moleculares associado ao método de BSA (Bulked Segregant Analysis). O polimorfismo entre os grupos foi geneticamente relacionado ao loco que determina a característica de resistência ou sucetibilidade. Dentre os 720 primers testados, 19 foram polimórficos, porém, apenas o marcador AK 01 manteve-se presente, quando testado em todos os indivíduos da população, mostrando-se a uma distância genética estimada de 20 cM em repulsão ao gene de resistência.

Avaliação do polimorfismo no gene da metilenotetrahidrofolato redutase e concentração de folato e vitamina B12 em pacientes portadores do HIV-1 em tratamento com anti-retrovirais

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Caracterização da resistência do algodoeiro a Ramularia areola e variabilidade molecular do patógeno

This research was conducted with the aim to study the genetic and pathogenic structure of Ramularia areola isolates collected in Brazil and to characterize the resistance response in cotton plants to ramularia spot. The genetic variability of 28 isolates of R. areola was studied using RAPD markers. The pathogenicity evaluation was realized by the inoculation of 6 isolates on cotton varieties Guazuncho-2 (Gossypium hirsutum) and VH8-4602 (Gossypium barbadense). The inheritance of disease resistance was studied using an artificially inoculated population of F2 individuals derived from the intercross of Guazuncho-2 (susceptible variety) end VH8-4602 (resistant variety), and also the parents and F1 individuals. Molecular polymorphism between the G. hisutum varieties DeltaOpal (suscetible) and CNPA CO-11612 (resistant) was estimated by 118 SSR and 24 AFLP markers. The parental genotypes Guazuncho-2 and VH8-4602 were selected for mapping, and then Recombinant Inbred Lines (RIL´s) derived from this crossing were evaluated with SSR 12 markers. The analysis of population structure of R. areola revealed that the three subpopulations were genetically simillar (Gst=0.18), and the isolates from Goiás and Minas Gerais were more similar to each other (0,92). This probability can be related to the relatively high gene flow among the three subpopulations (Nm=2.20). The isolates R. areola 9.1, from Minas Gerais State and 8.1 and 8.3 from Goiás State were the most aggressive ones to the susceptible variety Guazuncho-2. The variety VH8-4602 presented high level of resistance to ramularia spot. No differential interaction was observed between the pathogens and the analyzed varieties, and the resistance was classified as horizontal. The quantification of disease by number of necrotic lesions and number of spores in individual plants of F1 and F2 generations from the crossing between the varieties Guazuncho-2 and VH8-4602 presented continuous distribution, suggesting polygenic resistance. The resistance is probabilly recessive, since necrotic lesions and sporulation were observed on F1 plants. The molecular polymorphism between DeltaOpal e CNPA CO-11612 lineages was low (6%), then would be difficult to accomplish molecular mapping of disease resistance using this intercross. With the genotyping of the RIL s it was verified that 25% of the markers segregated in the proportions proposed by Mendel s Law and 75% of the studied markers presented segregation distortion in favor to the parental G. hirsutum. Both the low genetic variability of the pathogen and the number of resistance genes suggest that durable genetic resitance may be achieved

Caracterização genética de vacas leiteiras por meio de marcadores moleculares e suas implicações na composição e qualidade do leite

The objective of this study was to identify DNA polymorphisms at the genes leptin, β-lactoglobulin and pituitary-specific transcription factor in three genetic groups of Holstein x Guzerat dairy cows and investigate the relationship between their genotypes and the composition and quality of milk of dairy cows. Samples were collected in August 2009, being 113 blood samples from lactating crossbred cows and 58 milk samples. For analysis of DNA polymorphisms blood samples were collected, analyzed later in the Genetic Laboratory affiliated to the Zootechny Institute of São Paulo and individual milk samples were collected according to standards established by the laboratory of Management Program of Northeast Dairy Herds (PROGEN), at Federal Rural University of Pernambuco (UFRPE) for analysis of milk composition and quality. The characterization of genotypes was performed by PCR-RFLP, for which were designed specific primers for each studied gene and restriction enzymes Kpn2I, HaeIII and HinfI that cut the DNA of the following genes: leptin, β-lactoglobulin and a PIT, respectively. The leptin estimate genotypic frequence were CC 0.112, TT 0.225 and CT 0.661, for β-lactoglobulin were AA 0.136, AB 0.323 and BB 0.539, and for PIT were ++ 0.655, -- 0.311 and +- 0.032. The results show that the population is in Hardy-Weinberg disequilibrium for leptin, β-lactoglobulin and a PIT due to excess of heterozygotes in the population, however, as these genes are associated with the milk production it is considered that the animals have genetic potential for milk production in the Brazilian semi-arid conditions. Through the characterization of the studied herd there were not found implications of the polymorphism of leptin, β-lactoglobulin and PIT in the composition and quality of milk from cows in the different genetic groups 1/2, 3/4 and 7/8 Holstein x Guzerat. Key words: β-lactoglobulin, crossbred cows, leptin, PCR-RFLP, PIT1, semi-arid.

Ganho genético com base no tamanho efetivo populacional de progênies de seringueira

O objetivo deste trabalho foi estimar os ganhos genéticos de um teste de progênies de seringueira para a produção de borracha seca e, com base no maior tamanho efetivo populacional e maior ganho genético, obter os melhores indivíduos. Foram utilizadas 30 progênies de meios-irmãos, provenientes de sementes de polinização mista - alogamia e autogamia - de testes clonais no Estado de São Paulo. Utilizou-se o delineamento experimental de blocos ao acaso, com 30 tratamentos (progênies), 3 repetições e parcelas lineares de 10 plantas, em um espaçamento de 3x3 m, o que totalizou 900 plantas úteis. Aos três anos, o perímetro, a 50 cm do solo (PA50), e a produção de borracha seca (PBS) foram avaliadas por meio do teste precoce de produção Hamaker Morris-Mann (HMM). As variáveis foram analisadas pelo método de modelo linear misto, via procedimento REML/BLUP, em progênies com sistema reprodutivo misto e taxa de autofecundação de 22%. A identificação dos 20 melhores indivíduos quanto à PBS e ao PA50 proporcionou ganho genético de 67,96 e 16,48%, respectivamente, e um coeficiente de endogamia de aproximadamente 2,82%. O teste de progênies proporciona produção de sementes com melhor valor genético, grande variabilidade e baixa endogamia

Um algoritmo genético especializado aplicado ao planejamento da expansão do sistema de transmissão com alocação de dispositivos de compensação série

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Characterization of gene polymorphisms related to immune system physiology in Mangalarga horses

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

Estudo de simulação de modelos lineares mistos com distribuição normal contaminada no melhoramento genético animal

Objetivou-se com esse trabalho comparar estimativas de componentes de variâncias obtidas por meio de modelos lineares mistos Gaussianos e Robustos, via Amostrador de Gibbs, em dados simulados. Foram simulados 50 arquivos de dados com 1.000 animais cada um, distribuídos em cinco gerações, em dois níveis de efeito fixo e três valores fenotípicos distintos para uma característica hipotética, com diferentes níveis de contaminação. Exceto para os dados sem contaminação, quando os modelos foram iguais, o modelo Robusto apresentou melhores estimativas da variância residual. As estimativas de herdabilidade foram semelhantes em todos os modelos, mas as análises de regressão mostraram que os valores genéticos preditos com uso do modelo Robusto foram mais próximos dos valores genéticos verdadeiros. Esses resultados sugerem que o modelo linear normal contaminado oferece uma alternativa flexível para estimação robusta em melhoramento genético animal.

Efeito do grupo genético sobre as características de carcaça e maciez da carne fresca e maturada de bovinos superprecoces

Objetivou-se estudar o efeito das diferentes proporções de sangue Simental e Nelore sobre as características da carcaça e da carne de bovinos superprecoces. Foram utilizados 72 bovinos jovens inteiros (18 Nelore; 18 ½ Simental × Nelore; 18 Simbrasil e 18 Simental), com 8 meses de idade e 250 kg PV médio inicial. Os animais foram desmamados aos 8 meses de idade em sistema creep-feeding e posteriormente confinados durante 150 dias até atingirem o peso de abate, acima de 465 kg, e abatidos em frigorífico comercial. Os valores de pH e temperatura durante o resfriamento das carcaças foi semelhante para todos os grupos genéticos. da mesma forma, as variáveis carcaça fria, dianteiro e traseiro, não apresentaram diferenças entre os grupos genéticos. Os cortes foram bastante homogêneos, com excessão do contrafilé e do filé-mignon, que foram maiores nos animais Simental. Os animais da raça Nelore e ½ Simental apresentaram maior força de cisalhamento (4,98 e 4,45 kgf) em relação aos Simental e Simbrasil (3,13 e 3,33 kgf). No entanto, após a maturação da carne durante sete dias, não se constataram diferenças entre os valores de maciez entre os grupos. As perdas por evaporação e gotejamento foram maiores na carne in natura para os animais Simental e Simbrasil, no entanto, aos sete dias de maturação se tornaram semelhantes. O sistema de produção de bovinos superprecoces produz carcaças e cortes semelhantes entre as diferentes raças estudadas. Aos sete dias de maturação, a maciez da carne de animais Nelore foi semelhante à dos demais grupos genéticos utilizados neste estudo.

Prevalência de dislipidemia em adultos de meia-idade com polimorfismo do gene NOS3 e baixa aptidão cardiorrespiratória

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