Scheduling of scientific workflows using Niched Pareto GA for Grids

In a grid computing environment, many resources (compute, data, I/O, instruments, etc) are involved to solve a single large problem that could not be performed on any one resource. It is possible that the job submission for the resource request by resource consumers can be large owing to wide area distribution of grid. Key services such as resource discovery, monitoring and scheduling are inherently more complicated in a grid environment. In this paper, we approach the problem of grid workload scheduling by employing a Niched Pareto based genetic algorithm (NPGA) to generate near to optimal solution. In addition, evaluation of other scheduling mechanisms like first come first serve (FCFS), earliest deadline first (EDF) are compared. The results reveal that the proposed Niched Pareto genetic algorithm performs well compared to the other scheduling mechanisms when considering the workflow completion within the deadline