A dynamic-balanced scheduler for Genetic Algorithms for Grid Computing

The new paradigm of distributed computation, grid computing, has given rise to a large amount of research on resource scheduling. Unlike the distributed computation, grid computing uses heterogeneous resources, for what grid computing entails new challenges as the adaptation of parallel algorithms before developed for homogeneous resources cluster to the dynamic and heterogeneous resources. In this paper we present a dynamic-balanced scheduler for grid computing that solves two typical kinds of problems of grid computing, using for them the cycles of some resources of the grid. The first problem is based on iterative tasks that usually appear in optimization problems. The second problem is a directed acyclic graph (DAG) problem. Experimental results using dynamic-balanced scheduler show that it is possible to obtain an improved use of the resources in the grid. This strategy enables to adapt the length of a task to the computing capacity of each resource at any given moment. Furthermore, this scheduling strategy enables to execute all the tasks in a shorter time. Key-Words: Grid computing, dynamic-balanced scheduler, genetic algorithm, optimization problem.