Approaches Of Business System Model Repair Based On Change Mining

Today,with the development of big data technology,the role of business process management and optimization in enterprises is becoming more and more significant.However,as time goes by,business processes are constantly evolving,so that there are differences in behavior between logs and models,and the actual system.The log cannot be played back in the model.The reasons for the changes include factors such as regulations,policies,user needs,etc.,which have changed the business behavior of the system.As a result,it has become a hot topic in the field of process management to accurately and efficiently mine business process changes and repair business system models in a timely manner.Research on model repair mainly relies on the complete event log and the known situation of the model,while the process change analysis research based on incomplete event log is relatively rare.Existing research on process change analysis is mainly reflected in three aspects:one is the analysis of the change domain under the relationship between the behavior profile of the process model;the other is the propagation of multi-view changes based on control flow and data flow;and finally through log change detection.Discover or collect changes in the process management system.In order to make up for the limitations of existing methods,under the premise of incomplete logs,the influence of log behavior changes on model changes is detected,so that the results of change detection can be used to repair models and calculate the similarity between models.This paper mainly analyzes the changes of the detection process model based on the incomplete log behavior profile and repairs the model according to the relevant changes.The main research is as follows:(1)In order to obtain the process model based on the incomplete event log,the simulation tool CPN Tools can be used to conduct experiments.In the process of in-depth understanding of the simulation tool CPN Tools,in order to solve the lack of dynamic logic of the queuing system and its behavior analysis ability,a reasonable top-level CPN model was introduced,and the three main components of the queuing system were described in detail through the CPN model Part;Explains the powerfulness of the CPN Tools tool to establish a simulation model.(2)A change detection method based on incomplete logs is proposed.When the original model of the system is unknown,the activity comparison matrix(behavior contour matrix)is used to describe the Delete,Insert and Move change operations of the event log,so as to find that the original event log is the same as the system According to the deviation domain of the local change of the actual log,the model found in the original log is repaired according to the deviation domain of the log,and seven algorithms for realizing model repair are proposed.Finally,experiments are carried out on the proposed method through CPN Tools and ProM simulation tools,and the results verify the validity and correctness of the method.(3)A method for evaluating business process similarity with hidden changes is proposed.First,based on the known types of hidden transitions,the impact of various types of hidden transitions on the behavioral relationships between visible transitions is summarized,and the WI-TAR algorithm is proposed by using cover trees and a complete set of occurrence sequences to obtain the behaviors between related transitions.The relationship;then proposed the behavior contour similarity with hidden transition and the distance similarity calculation method based on the behavior contour;finally,the similarity algorithm in this paper is analyzed through experimental data.The experimental results show that this method can accurately calculate the similarity between models and has good operating efficiency.Figure[33]table[18]reference[84]...