On Optimization-Based Deadline Division for Workflow Scheduling

Complex applications are usually modeled as workflows, which are often represented by Directed Acyclic Graph (DAG) and require the power of Grid to run efficiently. Cost-optimized workflow scheduling under deadline constraints is a fundamental and intractable problem on Grids. In this paper, an effective and efficient heuristic for workflow scheduling is proposed. A mixed integer programming (MIP) approach is applied to find the optimal decomposition of global deadline constraint into time windows for all tasks. In terms of the time window allocations, the resources that minimize execution costs while satisfying these local time window constraints can be determined. The results of experimental evaluation show that our approach outperforms existing solutions in terms of workflow execution cost while meeting the deadline constraint.