Analysis And Optimization Of Business Process Change Region Based On Event Logs

The analysis of the change region of the process model aims to study the impact of the change operation on the model.The process model structure,the activity behavior and other aspects are analyzed.The study includes positioning the model change region,the impact of the change region on the model and the propagation of changes.Some existing researches based on the given target model adopt static analysis and use fixed observation points of the model structure to analyze the change region.However,such methods need the explicit target model and sometimes static analysis may not be able to locate the model change region quickly and accurately.Therefore,in this study,the method of analyzing the change region of the process model is based on the event logs and it gives a measure of the consistency distance of the causal behavior of the process model based on the event logs,the analysis method of change region of process model based on event logs,and the query method of process model based on event logs compliance measure.The part of the structure of the query result of the model is reused and it is merged into the newly established online shopping model.Then the new model is optimized.The object of business process change region analysis should be a process model pair with a certain degree of consistency.It is not of great significance to perform change region analysis on two completely inconsistent models.The target model of change region analysis may not be directly given,for example,it is given in the form of the event logs.And the impact of changes can not be ignored while locating the change region.Change region analysis can help people to select the best model in the model library to reuse models.It needs to quantify the model query index.And the reference objects used to select the model may not be given in the form of model structure fragments.Finally,reusing part of the model structure requires rebuilding the model to achieve the purpose of complete model and functions.Modeling process should consider issues such as multi-system interaction and model optimization.The main contributions of this article are as follows:(1)Aiming at solving the problem of measuring consistency only from the perspective of model structure.A method of measuring consistency of business process models based on complete event logs is proposed,which overcomes the situation that the target model is not directly given but presented in the form of complete event logs,and consistency analysis is from the two dimensions of model structure and behavior profile.Through the log-based behavioral profile analysis of the target model,the causal behavioral profile consistency and model distance from the source model are calculated,and then,the causal behavioral consistency distance between the models is calculated.Finally,the online shopping system is used as an example for verification.(2)Aiming at solving the problem that the target model is not directly given,a change region analysis method based on the complete event logs of the target model is proposed.Through a full analysis of the complete logs of the target model,the functional blocks of the process model and each activity path are divided,and the minimum k-order succession relationship between the source model and the target model for the strain transition is listed,and then the closeness of each activity path and the source model is measured.Based on the correspondence between the log-based behavior profile and model structure,the change region of the target model and the activities and paths affected by the change are determined.Finally,the online shopping system is used as an example to verify.(3)Aiming at the problem that the model query reference object is not a model structure fragment,this paper proposes a model query method based on the event logs compliance measure.Firstly,the behavior profile of the model is analysed,then the compliance of the query logs and each candidate model is calculated to quantify the query index,and the execution order compliance,enforcement compliance and causal coupling compliance are respectively calculated because of the incompleteness of the logs.Then,the group log compliance degree is obtained.By comparing the numerical values,it can select the best applicable model.Finally,the online shopping system is used as an example to verify.(4)Optimization and reuse of online shopping process modeling,reusing part of the model structure of the model query results in the previous chapter,and integrating it into the new model.The establishment of the new model uses the theory of Petri nets and behavioral profile,the behavioral interaction between buyers,sellers,and logistics systems is considered,and the exchange problem is optimized,and finally it adds configuration information to make the model more reasonable and effective.Figure[14]Table[13]Reference[89].