Comparação algoritmos de escalonamento de tarefas para grades computacionais sob diferentes métricas

Abstract

Grid computing, although advantageous in many scenarios, has as its greatest challenge the scheduling of different tasks in a heterogeneous system of machines that undergo constant changes in the availability of computational resources. Therefore, being able to compare the different heuristics, their results and differences under various circumstances is indispensable to select the best method for ordering the tasks. In view of this frequent problem in academic discussions, this work intends to analyze some of the classical heuristics for scheduling tasks by improving and running a simulator that implements them in different scenarios, extracting metrics for comparison and facilitating future implementations of new heuristics.