A combined approach for cloud tasks scheduling based on NSGA-II and harmony search

Abstract

Abstract The task scheduling is to map and arrange the execution of tasks on resources to optimize one or more efficient criteria. This paper tries to provide an improved model for scheduling workflow tasks of the cloud, which simultaneously considers three aspects of optimizing: makespan time, resource utilization, and scheduling length ratio. The model combines the non-dominated sorting genetic algorithm and the harmony search approach to achieve these goals. Harmony search algorithm attempts to perform a local search around the best solutions in each repetition of the algorithm to prevent it from getting stuck in the local optimal. The results on the four Cybershake, Epigenomics, Inspiral, and Montage datasets depict that the suggested algorithm is more efficient for all three criteria.