A GA based job scheduling strategy for computational grid

To meet the ever increasing demand of computing capacity and to cater towards collaborative engineering grid systems have emerged as the effective solution. Scheduling on such systems play a very critical role to ensure that the available resources are assigned and used optimally. Scheduling on grid has been proven to be an NP hard problem owing to the heterogeneity of the participating resources, huge search space and dynamicity of the grid system. GA has gained popularity among the researchers for scheduling problems on the grid systems, being a part of the evolutionary computing based on Darwin´s theory of survival of the fittest. The proposed work presents a design and eventual analysis of a scheduling strategy using GA that schedules the job with the objective of minimizing the turnaround time of the job. This is done ensuring the resource assignment as per the job´s requirement while considering the precedence and communication cost constraints. The work explores the use of Roulette wheel and Rank selection method as the selection methods while varying other parameters to study the model under various conditions.