Skip to Main Content
Many efforts have been made in the last few years to solve the high-level scheduling problem in Grid computing, i.e., the efficient resources utilization and allocation of workload within resources domains. Nowadays, some trends are based on the consideration of Fuzzy Rule-Based Systems, whose performance is critically conditioned to theirs knowledge bases quality. In this sense, Genetic Algorithms have been extensively used to obtain such knowledge bases, mainly founded on Pittsburgh approach. However, new strategies are recently emerging showing improvement over genetic-based learning methods. In this work, comparative results of two non-genetic learning strategies derived from bio-inspired algorithms, Differential Evolution and Particle Swarm Optimization, are presented for the evolution of fuzzy rule-based meta-schedulers in Grid computing.