Fault Tolerance in Grid using Ant colony Optimization and Directed Acyclic Graph
Vahid Modiri, Morteza Analoui, Sam Jabbehdari · International Journal of Grid Computing & Applications · 2011
By day to day developing the grid systems, it is necessary to apply new methods for allocating the resources to achieving the high performance in heterogeneous computing environment. This paper aims to seeking for a new approach by which one can allocate the tasks using a modified version of Ant Colony Optimization algorithm such that this algorithm cannot be involved in local minimum. Tasks will be entered to the system by Directed Acyclic Graph (DAG). This method tried to allocate the tasks to the processors in short time such that the tolerability of system may be considerably increased against faults. Any allocation stage comprised from two different phases and for any phase there has been provided new heuristic function. In addition, there will be introduced a new mechanism for updating the ant’s pheromone. The results of simulation using JAVA programming language indicates that such approach can be used most efficiently for problems of allocating the grid tasks.