基于贝叶斯理论的工作流任务分配模型的设计

Design of task assignment model for workflow based on Bayesian theory

在需团队协作完成的工作流应用中,不同类型任务需要经过多级的人来完成。但在实际运行过程中,随着任务历史数据的积累,可在现有数据前提下判定新来任务是否可以跳过固有业务流程。以贝叶斯网络为理论基础,结合移动广告系统审批流程的特点,对任务跨级分配的可行性建模分析,减少非必须的任务处理流程和任务在系统中的流转时间,从而提高工作流系统的吞吐量和任务的处理效率。

In the workflow application which needs team collaborative, tasks are completed by several people. However, in practical, whether inherent business processes can be skipped could be computed fi＇om existed historical data. This paper constructs a task assignment model to analyze the feasibility of tasks assignment of cross-level based on Bayesian network and utilize the characteristics of the approval task to reduce non-essential processes and the time of tasks flow in workflow system. As a result, improve throughput of system and efficiency of processing tasks.