Wpso: a Pso-Based Algorithm for Cost-Effective and Deadline-Constrained Workflow Scheduling in Cloud Environments

With the growing scale and complexity of scientific computing in the big data era, efficient scheduling of scientific workflows in cloud environments has become a significant research challenge. Scientific workflows, modeled as Directed Acyclic Graphs (DAGs), require efficient allocation of heterogeneous cloud resources to optimize execution cost while meeting deadline constraints. This paper proposes a novel PSO-based algorithm, WPSO, for workflow scheduling. WPSO introduces enhancements in encoding, decoding, and particle updating to address resource heterogeneity, complex cost models, and task dependency constraints. Experimental results on four realworld scientific workflows (CyberShake, Epigenomics, LIGO, and Montage) across various scales demonstrate WPSO's superior performance compared to existing algorithms in terms of success rate and normalized cost. The proposed method achieves significant improvements in resource utilization and cost-effectiveness under stringent deadline constraints, providing a robust solution for workflow scheduling in cloud environments.