Research On Similarity Of Business Process Based On Artifact Lifecycle

Increasingly complex process models in the enterprise process respository make it difficult for organizations to manage processes.Analyzing the common areas of different processes and determining the similarity of processes not only provides a theoretical basis for processes improvement,process consolidation,and process reusability,but also reduces the complexity of managing business processes.Most of the existing process similarity computation methods were modeling based on control flow,which neglects the consideration of business data.The Artifact-centric process places business data in process of execution at the core.This paper proposes a process similarity calculation method based on Artifact lifecycle for the problem of insufficient think of business data in the existing process similarity research method.The main contents of this paper are as follows:Firstly,the Artifact lifecycle is modeled to obtain the Artifact lifecycle tree.Calculate the edit distance and the valuation function of the attribute assignment sequence in the Artifact log instance,match the best attribute assignment path in the Artifact lifecycle tree.Calculate log and model attribute assignment sequence similarity and activity sequence similarity separately,and the measuring method of business process similarity based on Artifact lifecycle model is proposed.Secondly,the Petri net is used to formally define the Artifact lifecycle model and analyze the rule pattern of the Artifact attribute assignment order.By calculating the suffix dependency in the attribute assignment sequence similarity and the similarity of Artifact in the log,an Artifact-centric process log similarity measuring algorithm is proposed.Finally,the measuring method of business process similarity based on Artifact lifecycle model is realized by simulation data,and the practicability of the algorithm is effective.On the real data set,the Artifact-centric process log similarity measurement algorithm is implemented and compared with the TAR algorithm to prove the accuracy and efficiency of the algorithm.