Similarity Measure Method For Artifact-Centric Business Process

With the arrival times of service economy,enterprises are facing the keen competition environment.Thus,enterprises have to adjust their business pattern and management way.As the core business assets of enterprise,the number and complexity of business process is increasing.To manage business processes become more difficult for enterprises.In order to effectively manage and use the existing business processes,process retrieval,process integration,and process clustering have gotten more and more attention from academia and industry.The key here is to judge the similarity between business processes.Existing process similarity computation methods mostly focus on the process-centric business process model,less consider to the data in the process.However,the data involved some business process is essential to help developers to understand and implement the business process.Recently,researchers present some data centric business process models to represent all kinds of data,for instance,BP schemas,business data,and runtime status and dependencies.Due to Artifact models lead this trend by using an information model for data in a BP and a lifecycle model to capture how the business data evolve through business operations in the business process,the Artifact-centric business process model is a widely accepted data-centric business process.This paper starts with three aspects of data state,data flow and data operation sequence,and puts forward three novel similarity measuring method for Artifact-centric business process.The main contents of this paper are as follows:(1)Firstly,through the study of the state change of the data in the Artifact lifecycle,we propose an Artifact-centric business process similarity measuring method based on state change of the data with adjacency relations.The method uses the data state to describe the operation of the task to the data,and measures the similarity of state change of the data with adjacency relations in the Artifact lifecycle to evaluate the similarity of Artifact-centric business process model.The experimental results show that the proposed method is superior to the existing methods based on task adjacency relations.(2)Secondly,an Artifact-centric business process similarity measuring method based on task context is proposed.In this method,the dependency between tasks is designed as a weight function,the same data with the same operation and task input and output data ratio as the value function,using the product of two functions to evaluate the similarity between two Artifact-centric business process.Experiments show that this method can solve the problem of similarity measurement of Artifact-centric business process,and make up for the lack of consideration of Artifact data based on task dependency.(3)Thirdly,a novel Artifact-centric business process behavior similarity measuring method is presented.The operation sequences of Artifact data attribute are extracted from expanded Artifact-centric business process model.Then,Data operation sequences are classified,and the similarity of different operation sequence is calculated.Lastly,the similarity of all classes of data items was aggregated to measure the similarity between Artifact-centric business processes.The results showed that the proposed method is not only more consistent with the properties of process behavior similarity algorithm than the existing algorithm,but also a good similarity measuring for data-centric business process.(4)Finally,the effectiveness and efficiency of the similarity measuring methods are validated through experiments by using the real data set of Hangzhou Zhongfang Information Co.,Ltd.The experimental results show that all methods we proposed possess all the properties of similarity measuring method based on business process behavior,and support the Artifact-centric business process model similarity measurement.