Abstract
Background: Task scheduling problems are very important issues in cloud computing ecosystems because a scheduling plan determines which tasks should be executed on what resources. Therefore, the task scheduling plan for execution has deep effects on the final solutions. As the total execution time, or so-called makespan, is the most prominent performance assessment metric that users endure. Therefore, the goal is to schedule tasks for execution so that the makespan is minimized that is an NP-hard problem.
Method: Most of existing literature methods algorithms suffer from a local optimal trap. To fill the gap, a hybrid shuffled frog leaping algorithm (HSFA) is proposed to solve task scheduling problems in cloud environments. The novelty of the current paper is to design a Hill-climbing algorithm that is customized and randomly called to possibly enhance the gained solutions of the exploration phase for strengthening the exploitation phase.
Results:To verify the effectiveness of the proposed HSFA in solving task scheduling, numerous scenarios are conducted. The simulations in the Python programming environment prove that HSFA outperforms against GA, GWO, and canonical SFA in terms of makespan minimization significantly.
Conclusion: The proposed hybrid HSFA algorithm strikes local and global searches efficiently which gives promising results. In addition, this algorithm has better performance running on communication-intensive DAG applications in terms of makespan minimization in comparison with other counterparts.