Cost optimization for workflow applications with
deadline constraints is fundamental and intractable in
Grids. In this paper, early tree is introduced to find an
early feasible schedule for a workflow application.
According to the early tree, a cost optimization
algorithm is proposed. Taking into account the
workflow total float, the workflow deadline is
segmented to activity deadlines while keeping
precedence constraints. Costs of all activities are
locally optimized, so does the workflow cost.
Experimental results show that the proposal that this
approach can dramatically decrease workflow cost with
different deadlines. Moreover, it outperforms other two
leveling algorithms in performance on average.