基于Petri网可达图的业务对齐方法

Business alignments based on reachable graphs of Petri nets

为提高过程挖掘中计算最优对齐的效率,提出一种基于Petri网可达图的业务对齐方法。首先,根据给定事件日志,提取其包含的活动子集;然后,将子集中包含的活动映射到变迁,构建花型日志模型;进一步,计算花型日志模型与过程模型之间的乘积模型及其可达图;最后,给出算法在可达图中查找出事件日志中全部迹与过程模型之间基于给定代价函数的一个最优对齐和所有最优对齐。从理论上证明了该方法的适用性与有效性。通过仿真实验,验证了该方法的可行性与优越性。

To improve the efficiency of alignments in the process mining,a new alignment method was presented between event logs and process models based on reachable graphs of Petri nets.A subset was generated including all the activities in the given event log.The flower log model was built according to all the activities in the subset.The product of the flower log model and the process model was calculated,as well as its reachable graph.On the basis of the reachable graph,two detailed algorithms were presented to obtain an optimal alignment and all optimal alignments between the trace and the model based on the given cost function respectively.The availability and effectiveness of the proposed method were proved theoretically.After a series of simulation experiments,the feasibility and superiority of the proposed method was verified.