Business Process Change Mining And Optimization Analysis Induced By Log

In recent years,more and more attention has been paid to business process management in enterprises or organizations,and the cost of enterprise operation has been greatly saved with the help of business process model.Due to the complex and ever-changing needs of users and constantly updating technical requirements,the business process model that enterprises run with it also needs to be constantly changed.In the business process management,the change mining technology can be used to accurately locate the changed activity set in the system,so as to optimize the reasonable configuration of the model and make it more in line with the actual needs.Therefore,this thesis proposes a log-induced business process change mining method and optimization analysis has a certain theoretical significance and practical value.At present,most of the research on change mining is to compare and analyze the behavior profile from the point of view of the model.However,with the arrival of the era of big data,business processes become more and more complex,and model-based mining changes become extremely difficult,so mining system changes from the perspective of event logs has increasingly become the focus of people's research.On the one hand,this thesis excavates the changes in the system based on the event log using the idea of conditional behavior probability and optimal alignment,on the other hand,it looks for the change domain of the merged model from the process tree and sequence matrix.Finally,the configuration optimization of the model with behavior constraints is analyzed.The main contributions of this thesis are as follows:(1)In order to solve the problem of system changes in business process model mining,the existing methods mainly mine changes when both the source model and the target model are known,but seldom consider the situation when the target model is unknown.A change mining method of aimless model based on event log is proposed.Starting from calculating the compliance of the event log and the source model,the log is preprocessed,and on this basis,the behavior relationship between each pair of activities is obtained by calculating the conditional occurrence probability of the activity pairs in the log.Thus,the changed activity set in the system can be found by comparing with the behavior profile in the source model.This method analyzes the event log from the dynamic change of behavior,and effectively overcomes the limitations of the existing methods(2)In order to solve the problem of event log mining changes generated by the system,the existing literature mainly studies the changes of some activities in the log,but pays little attention to the changes of the whole event trajectory in the log.A change mining method based on optimal alignment of cost function is proposed.From the point of view of calculating the high-frequency morphological fragments in the log,the optimal alignment of each trace is found,and on this basis,the similarity between each trace in the change log and the source log is calculated,and then the changed trace is obtained.This method can effectively mine the trace set of changes in the system and ensure the integrity of the changes.(3)In order to solve the problem of the change domain of the merged business process model,the existing methods mainly study the inconsistent regions during the interaction of the model,but seldom consider the change domain of the configurable process model after the merger.A change domain analysis method of configurable business process model based on process tree is proposed.Firstly,the configurable process model is obtained by merging the source model from the point of view of the process tree,and the existence of the change domain is determined by calculating the average weight distance between the model and the source model.Then the change domain of the configurable process model can be accurately located by sequential matrix.This method solves the problem of merging model change domain quickly and efficiently,and lays the foundation for the following configuration optimization(4)Aiming at the problem of business process model optimization,the existing literature mainly studies the optimization of business process model by adding or deleting some transition activities in Petri net system,and pays little attention to the optimization analysis method of configuration constraints.An optimization analysis method of Mobile Phone Top-Up Service system based on configuration constraints is proposed.Starting with the specific example of Mobile Phone Top-Up Service,combined with the complex and changeable user needs,the model with behavior constraints is optimized by adding corresponding blocking transition and hidden transition,and the unconstrained objective model is obtained.This method effectively improves the correctness of the business process model after configuration.Figure[21]table[22]reference[82]...