Mutation and dynamic objective-based farmland fertility algorithm for workflow scheduling in the cloud

• Published in: Journal of parallel and distributed computing Vol. 164; pp. 69 - 82
• Main Authors: Li, Huifang, Wang, Yizhu, Huang, Jingwei, Fan, Yushun
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.06.2022
• Subjects: Cloud computing; Deadline constraints; Meta-heuristics; Workflow scheduling; Cloud computing; Deadline constraints; Workflow scheduling; Meta-heuristics
• ISSN: 0743-7315; 1096-0848
• DOI: 10.1016/j.jpdc.2022.02.005

Nowadays, many scientific applications are deployed in the cloud to execute at a lower cost. However, the growing scale of workflows makes scheduling problems challenging. To minimize the workflow execution cost under deadline constraints, this article proposes a Mutation and Dynamic Objective-based Farmland Fertility (MDO-FF) algorithm for obtaining a near-optimal solution within a relatively shorter time. A Dynamic Objective Strategy (DOS) is introduced to accelerate the convergence speed, while a multi-swarm evolutionary approach and mutation strategies are incorporated to enhance the search diversity and help to escape from local optima. By seeking new potential solutions and searching in its corresponding neighborhoods, our proposed MDO-FF can make a good trade-off between exploration and exploitation. Extensive experiments are conducted on well-known scientific workflows with different types and sizes. The experimental results demonstrate that in most cases, our MDO-FF outperforms the existing algorithms in terms of constraint satisfiability and solution quality. •Applying the Farmland Fertility algorithm to address workflow scheduling problems.•Introducing a mutation operator to modify the population updating mechanism.•Developing a multi-swarm evolutionary approach with different searching strategies.•Adopting a Dynamic Objective Strategy to efficiently search for feasible solutions.