An Opposition‐Based Chaotic Enhanced African Vulture Algorithm for Efficient Scientific Workflow Scheduling in Cloud

ABSTRACT With the increasing compute and data demands of intricate scientific applications, cloud computing has emerged as a key solution for executing scientific workflows while adhering to Quality of Service (QoS) parameters. Developing an efficient and cost‐effective solution, especially for large‐scale applications, continues to be an immensely challenging task. To handle this challenge, this research presents a scheduling algorithm, an opposition‐based chaotic enhanced African vulture workflow scheduler (OCEAVWS), to reduce the makespan and financial cost incurred in executing scientific workflows. The proposed approach initially employs opposition‐based learning (OBL) techniques to create a high‐quality initial population, utilizes a logistic chaotic map that assists in selecting phases in the African vulture optimization algorithm (AVOA), and is further enhanced by dimension learning hunting (DLH). The strength of the proposed method is assessed on the WorkflowSim tool on four different scientific workflows. The experimental results show substantial improvements, including an 18.03% reduction in makespan and an 8.07% decrease in financial costs. The evaluation metrics indicate that our approach achieves lower average relative deviation index (ARDI) values in all cases, lower S‐metric values in 83% of scenarios, and higher hypervolume in 92% of the evaluated scenarios. Statistical validation using Wilcoxon and Friedman tests confirms that OCEAVWS significantly outperforms five other state‐of‐the‐art multi‐objective optimization approaches, demonstrating its effectiveness in scientific workflow scheduling.