Using a Classifier System in a Dynamic Changing Environment:An Application to Dynamic Load Balancing

Jan CorreaAlba C. Melo

One important problem is to improve the performance of a network of computers when perfoming scientific computation. There are few attemps to apply the adaptation of GAs to the dynamic processor scheduling, a real-life problem that the algorithm has no prior knowledge of what will happen in the system and has no adapt quickly to enviroments changes. In this paper we propose a Classifier System that is a adaptative system that applies GA over a population of decision rules to achieve better decisions about process assignment. The results have been impressive and the classifier system was able to learn in a dynamic enviroment and to surpass the performance of a well human designed analytic criterion. Clique no link abaixo para buscar o texto completo deste trabalho na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: