Navegando por Assunto "Evolutionary computing"

Agora exibindo 1 - 2 de 2

Acesso aberto (Open Access)
Desempenho do algoritmo genético com iteração retroviral para otimização de funções com representação real
(Universidade Federal do Pará, 2015-06-30) FRANCO, Dielle da Silva Corrêa; SANTANA, Ádamo Lima de; http://lattes.cnpq.br/4073088744952858; OLIVEIRA, Roberto Célio Limão de; http://lattes.cnpq.br/4497607460894318
Viral Infection is used to improve the performance in Genetic Algorithms (GA) by reducing premature convergence through the population diversity control, since viruses presents high replication and mutation rates in the nature. The metaheuristic called AGRI is inspired biologicaly in a viruses family based on RNA, which provide a high allelic variation to GA, since RNA doesn’t have genoma correction mechanisms to remove re-combined viral genetic material . In this algorithm, the viruses are a separate population. To each infection, the better performance viruses genomes are transmitted vertically spreading parts of solutions to GA population. The diversity viral is maintained through a mechanism that substitutes all viruses out of elitism viral rate. In this method, the virus population evolves along with GA population, so the inefficient viruses are created from genetic material of the better adapted individuals and other new genes. The algorithm AGRI follows biological principles in several viral infection and multiplication aspects. For example: it creates the first viral population without GA population genetic material; it sorts the viral population before infect an individual, making possible some viruses doesn’t infected a part of the population and other viruses infect more individuals. Since GA second-generation, the replaced viruses are created by both individuals genetic material and have different genes quantities. In this approach, the search space maximization is increased by three mechanisms: high viral population genetic variability by variety of sizes to solutions pieces; infection validation process that confirms the fitness increases in each individual and infection possibility by any viruses in the viral population. To analyse the AGRI’s viral infection parameters effects and comparate his performance with others high-performing metaheuristics, the following minimization benchmarking are selected: F1 (Shifted Sphere Function), F2 (Shifted Schwefel’s Problem), F3 (Shifted Rotated High Conditioned Elliptic Function) e F5 (Schwefel’s Problem 2.6 with Global Optimum on Bounds). The results to the functions unimodais proposed showed that AGRI has a good performance in comparison with others metaheuristics reaching in few iterations the global best or good results.
Acesso aberto (Open Access)
Estratégias evolucionárias para otimização no tratamento de dados ausentes por imputação múltipla de dados
(Universidade Federal do Pará, 2016-02-16) LOBATO, Fábio Manoel França; SANTANA, Ádamo Lima de; http://lattes.cnpq.br/4073088744952858
The data analysis process includes information acquisition and organization in order to obtain knowledge from them, bringing scientific advances in various fields, as well as providing competitive advantages to corporations. In this context, an ubiquitous problem in the area deserves attention, the missing data, since most of the data analysis techniques can not deal satisfactorily with this problem, which negatively impacts the final results. In order to avoid the harmful effects of missing data, several studies have been proposed in the areas of statistical analysis and machine learning, especially the study of Multiple Data Imputation, which consists in the missing data substitution by plausible values. This methodology can be seen as a combinatorial optimization problem, where the goal is to find candidate values to substitute the missing ones in order to reduce the bias imposed by this issue. Metaheuristics, in particular, methods based in evolutionary computing have been successfully applied in combinatorial optimization problems. Despite the recent advances in this area, it is perceived some shortcomings in the modeling of imputation methods based on evolutionary computing. Aiming to fill these gaps in the literature, this thesis presents a description of multiple data imputation as a combinatorial optimization problem and proposes imputation methods based on evolutionary computing. In addition, due to the limitations found in the methods presented in the recent literature, and the necessity of adoption of different evaluation measures to assess the imputation methods performance, a multi-objective genetic algorithm for data imputation in pattern classification context is also proposed. This method proves to be flexible regarding to data types and avoid the complete-case analysis. Because the flexibility of the proposed approach, it is also possible to use it in other scenarios such as the unsupervised learning, multi-label classification and time series analysis.