Workflow Cost Optimization Heuristics Based on Advanced Priority Rule

The time and cost trade-off problem in workflow scheduling with deadline constraint in utility grids is an NP-hard, but attractive problem. In the past studies, the priority rule, best fit based on time-dependent coupling strength (referred as BFTCS), was adopted to solve this question because BFTCS considers both the features of resources and workflow structures and uses them as two important factors in the iterative heuristic for cost optimization. However, BFTCS did not consider another important factor: the temporal feature of the workflow. In this paper, we define temporal mobility (TM) to describe the temporal feature of the workflow, and incorporate TM into the iterative stage as an important factor to improve the workflow execution cost further. This novel priority rule (referred as best fit based on time-dependent coupling strength and temporal mobility, BFTCSTM) gives priority to the tasks with larger TM and smaller TCS (time-dependent coupling strength) during the iterative stage. Therefore, compared with BFTCS, BFTCSTM slows down the increasing rate of the workflow execution time greatly, and provides more opportunities to other tasks to optimize cost, which decreases the total cost further. The experimental results demonstrate the advantages of BFTCSTM based iterative heuristic.