基于搜索树的业务流程Petri网模型抽象化简方法

由于大型复杂的业务流程模型不便于用户形成快速的梗概理解,因此模型的抽象化简方法研究至关重要.已有的模型抽象化简方法大都考虑模型的控制流,很少考虑模型的交互语义,并且对于模型的待抽象区域定位模糊.本文利用Petri网的行为关系理论,结合深度优先搜索的思想,提出了基于搜索树的业务流程模型抽象化简方法.通过定义工作流网的3种块结构和边界库所的概念,得到网系统的变迁关联搜索树,进而利用变迁关联搜索树来识别模型中待抽象的区域,从而实现模型的抽象化简.最后,通过一个具体的实例来验证所提出方法的有效性和可行性.本文的主要贡献是:在同时考虑模型行为交互语义和控制流依赖关系的基础上,提出块结构的抽象化简方法和块结构的识别定位方法,算法的时间复杂度控制在多项式难度.

A search-tree-based abstraction method for business process Petri nets models

As it is difficult to form a quick overview understanding for large and complex business process models, so the studies about the technologies and methods of model abstraction and simplification are crucial. The related researches in existence have some limitations, such as only the control flows of the model are taken into account, whereas the interactive behavioral semantics are mostly neglected, and have little vague about the scopes orientation that to be abstracted in the model. A search-tree-based abstraction method for business process model is purposed in the paper, which is founded on the basis of behavioral relation theory of Petri nets and the Depth-First Search ideas. Firstly, the concepts of three kinds of block structures and boundary places in work-flow Petri nets are formalized, and the transitions association tree of the system is then obtained. The transition association tree is further used to identify and locate the areas that to be abstracted in the model, and the aim of model abstraction is then accomplished. Finally, a case example is applied to illustrate the validity and feasibility of the proposed method. Therefore, compared to the existing work, our main contributions can be stated as follows: we purpose a block based abstraction method and its corresponding block identification method, where the interactive semantics and control dependencies are both taken into consideration, and the purposed methods are in polynomial time complexity.