业务过程模型的重构技术研究

How to improve the quality of business process models is an urgent problem to be solved in the field of process aware information systems. However, the existing literature has made a lot of achievements in the quality problems of business process models, i.e. activity label quality, understandability, maintainability and structuredness, but less attention to efficiency and complexity. In this proposal, we propose refactoring techniques of business process models based on model efficiency and model complexity. Two key issues are explored in this proposal: (1) We propose a refactoring technique to mine concurrency of business process models. The technique can automatically identify false sequence relations between two adjacent tasks and false transitive sequence relations between two non adjacent tasks in business process models, change these false sequence relations to concurrency relations, and change these false transitive sequence relations to concurrency relations. This refactoring process establishes the dependency-isomorphism between the original model and the refactored one from the perspective of data to make sure that we can improve the efficiency of this original model without changing its data processing result. (2) We propose a refactoring technique to reduce structural redundancy of business process models. The technique can automatically identify implicit places in business process model using structure-based analysis approach, and design refactoring operations to eliminate these implicit places. This refactoring process establishes the basic relation-isomorphism between the original model and the refactored one from the perspective of control-flow to make sure that we can reduce the complexity of this original model without changing its behavior. The research results will provide new ideas and technical support for the model quality problems in process ware information systems.

如何提高业务过程模型的质量，是过程感知信息系统迫切需要解决的问题。现有业务过程模型重构研究，主要关注从活动标签质量、可理解性和可维护性方面提高模型质量，较少关注效率和复杂性。本项目从效率和复杂性视角，研究提高模型质量的业务过程模型重构技术，解决以下问题：（1）提出一种挖掘并发性的重构技术，自动识别业务过程模型中紧邻任务间的假顺序关系和非紧邻任务间的假传递顺序关系，并把这两种关系重构为并发关系，以提高模型的效率。该重构过程从数据视角把不改变模型的数据处理结果规约为建立重构前后模型间的相关性同构。（2）提出一种消除结构冗余的重构技术，使用基于结构的分析方法，高效地自动识别业务过程模型中的隐式库所，并设计重构操作消除隐式库所，以减少模型的复杂性。该重构过程从控制流视角把不改变模型的行为规约为建立重构前后模型间的基本关系同构。为企业实施过程感知信息系统面临的模型质量问题提供新的思路和技术支持。

如何提高业务过程模型的质量，是过程感知信息系统迫切需要解决的问题。本项目关注提高模型质量的业务过程模型重构技术研究，提出了：1）一种挖掘并发性的重构技术，在不改变业务过程模型外部行为的前提下，可以将紧邻任务间的假顺序关系和非紧邻任务间的假传递顺序关系重构为并发关系，提高过程模型的效率；2）一种消除结构冗余的重构技术，在不改变业务过程模型外部行为的前提下，可以删除业务过程模型中的隐式库所，减少过程模型的复杂性；3）一种挖掘业务过程模型的技术，可以从没有“aba”模式的日志文件中挖掘出含有2度循环的业务过程模型；4）一种确保业务过程正确协同的技术，可用于自动检测该过程模型是否是正确的。在论文方面，发表和接收学术论文30篇，其中，期刊论文22篇，会议论文8篇。期刊论文中，高被引论文2篇，CCF推荐A类中文期刊5篇；会议论文中，CCF推荐的B类会议2篇、C类会议1篇；在知识产权方面，授权发明软件1项，登记软件著作权4项；在人才培养方面，培养本科生12人、硕士生5人、博士生2人。