Aplicación de técnicas de inteligencia artificial en la solución del problema de ubicación en el diseño de bases de datos distribuidas
Archivos
Fecha
2006-06-24
Autores
Rosa Paz, Darién
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas
Resumen
El auge alcanzado por los sistemas de información distribuidos y el desarrollo de los manejadores de bases de datos comerciales, ha hecho que los sistemas de bases de datos distribuidas se conviertan en una alternativa viable para satisfacer las necesidades de información actuales.
La distribución de los datos es un problema crítico que afecta de manera considerable el desempeño global de los sistemas distribuidos, debido a que influye directamente en la eficiencia del procesamiento de las consultas. Debido a la complejidad del problema, las diferentes propuestas de solución presentadas hasta la fecha, han coincidido en dividir el proceso de diseño de la distribución en dos fases seriadas: la fragmentación y la ubicación de los fragmentos en los sitios de la red.
La presente investigación, titulada “Aplicación de técnicas de Inteligencia Artificial en la solución del problema de ubicación en el diseño de bases de datos distribuidas”, aborda el problema de la asignación de fragmentos en el diseño de bases de datos distribuidas, tomando como referencia el modelo matemático Özsu, al cual se le da solución aplicando las técnicas de Algoritmos Genéticos y el algoritmo Q-Learning del Aprendizaje Reforzado.
El producto final de la investigación es un asistente que será insertado en una herramienta general para el diseño de bases de datos distribuidas.
The boom reached by the distributed information systems and the development of the commercial data base managers, has made possible that the distributed data base systems become into a feasible alternative for satisfying the nowadays information needs. The data distribution is a critical problem which affects in a considerable way the overall performance of the distributed systems; given that it has a direct influence on the efficiency of the process of consulting. Due to the complexity of the problem, the several proposed solutions presented until today, have in common that they divide the process of designing the distribution into two successive phases: the fragmentation and the allocation of the fragments in the network sites. The present research, entitled “Application of Artificial Intelligence techniques to the solution of the allocation problem in the distributed data bases design”, deals with the problem of the assignation of fragments in the design of distributed data bases, taking as a reference the Özsu mathematical model, which is solved applying the techniques of Genetic Algorithms and the Q-Learning algorithm of the Reinforcement Learning. The final product of the research is an assistant which will be inserted into a general tool for the design of distributed data bases.
The boom reached by the distributed information systems and the development of the commercial data base managers, has made possible that the distributed data base systems become into a feasible alternative for satisfying the nowadays information needs. The data distribution is a critical problem which affects in a considerable way the overall performance of the distributed systems; given that it has a direct influence on the efficiency of the process of consulting. Due to the complexity of the problem, the several proposed solutions presented until today, have in common that they divide the process of designing the distribution into two successive phases: the fragmentation and the allocation of the fragments in the network sites. The present research, entitled “Application of Artificial Intelligence techniques to the solution of the allocation problem in the distributed data bases design”, deals with the problem of the assignation of fragments in the design of distributed data bases, taking as a reference the Özsu mathematical model, which is solved applying the techniques of Genetic Algorithms and the Q-Learning algorithm of the Reinforcement Learning. The final product of the research is an assistant which will be inserted into a general tool for the design of distributed data bases.
Descripción
Palabras clave
Algoritmos Genéticos, Aprendizaje Reforzado, Inteligencia Artificial, Bases de Datos Distribuidas, Diseño