Abstract: To improve the efficiency of conformance checking between event log and process model,a similar optimal alignment  method based on Petri net was proposed.According to the online shopping pattern,a Petri net model was built,and all of the optimal alignments between the given trace and models were calculated.The analysis showed that some optimal alignments had exactly the same set of movements with different ordering,which was defined as the similar optimal alignments.The properties of similar optimal alignments were analyzed,thus a series of theorems and corollaries were obtained.The similar relationship and equivalent relationship of optimal alignments was defined,and a partition of the optimal alignment set was given.A representative item was chosen to represent all the deviations.Through analyzing the four workflow patterns,MPA algorithm was presented to solve all the representatives of similar optimal alignments between models and restricted traces.The simulation examination was used to verify the correctness and effectiveness of proposed method.

Key words: trace, Petri net model, similar optimal alignments, workflow, alignment algorithm

摘要： 为了提高事件日志与过程模型之间一致性检查的效率,提出一种基于Petri网基本结构的相似最优校准计算方法。根据网上购物流程建立Petri网模型,计算给定迹与模型之间的所有最优校准。分析发现有些最优校准包含的移动集合完全相同,只是移动出现顺序不同,定义为相似最优校准。通过对相似最优校准的性质进行分析得出定理与推论。定义最优校准相似关系和等价关系,给出了最优校准集合的划分方法,该方法可选取代表项,体现迹与过程模型之间的所有偏差。通过分析四种工作流模式提出多阶段校准算法,用于求解Petri网模型与约束迹之间相似最优校准代表项。仿真实验例证了该相似最优校准方法的正确性和有效性。

关键词: 迹, Petri网模型, 相似最优校准, 工作流, 校准算法