Enhanced List-Driven Workflow Scheduling Optimization for Cloud Environments

The cloud environment has become a new platform for workflow scheduling. This paper presents a new list scheduling algorithm called the optimal predict priority task scheduling algorithm, which has two stages, task priority sorting and processor selection, and looks both up and down. In the task priority sorting stage, the improved pre-diction cost matrix considers not only the direct successor tasks' influence but also the direct predecessor tasks' influence on the pre-diction cost matrix. In the processor se-lection phase, the nonintersecting technique is used for processor selection. The experimental results on randomly generated application graphs and re-al-world application graphs show that IPPOSS outperforms the existing related algorithms.