Abstract: Aiming at optimizing accuracy in business process workflow represented by directed acyclic graph (DAG) in time constraint, the reverse reduction optimization algorithm (RRO) based on deadline was proposed. This algorithm gives up one-way target strategy of traditional algorithm and eliminates the problems of the low accuracy or high completion time of business process. By integrating the time debris left by the traditional algorithm and analyzing the relationship between the accuracy and the time of the service, the optimal path is obtained by reverse reduction. Simulation shows that the reverse reduction optimal algorithm can achieve the balance between completion time and completion accuracy in the range of deadline. Furthermore, the influence of the performance to the algorithm when it applies to different deadlines and different numbers of tasks is analyzed by different simulation data and the increasing of the deadline or the number of tasks improves the algorithm optimization performance.

Key words: workflow, time constraint, reverse iteration, accuracy optimization