The Research Of Cross-organizational Complex Business Process Mining Based On Event Log

With the rapid development of cross-organizational process-aware information system and the great rise of new Internet technologies characterized by "sharing",many challenging research topics have emerged in the field of process mining.In the "Process Mining Manifesto",IEEE Working Group on Process Mining pointed out that cross-organizational business process mining is one of the challenges in the field of process mining today,and two cases of cross-organizational business process mining are given:(1)the first case is based on mutual cooperation and collaborative cooperation among different organizations;(2)the second case is based on shared business process management infrastructures among different organizations.In this paper,two cases of cross organizational business process mining are studied.The main work and innovation are as follows:According to the cross-organizational business process mining of inter-cooperation and cooperation mode of different organizations,a distributed cross-organizational business process mining method is proposed.Taking the case of cross-organizational emergency business process as the research object,the formal definition of cross-organizational emergency business process tasks and the definition of the collaborative relationship among different tasks are given.The collaborative pattern between different organizations is defined according to the collaboration relationship between the tasks of the cross organization emergency business process.Then,the Petri net model based on extended message repository is used to describe the cross-organizational emergency business process model.The algorithm of mining private business process model and inter-organizational collaborative process model is given.Through the model integration algorithm,the private business process model in the organization and the inter-organizational collaborative process model are combined to obtain the global cross-organizational emergency business process model;Finally,the framework of mining the cross-organizational business process is given,and the Petri net editing tools are used to verify the correctness of the mined global cross-organizational business process model.Aiming at the cross organizational business process mining of shared business process management infrastructure,we study the distance between business processes of different organizations,and propose two computing methods for distance between business processes.(a)By the comparison of model to model,we give the method of calculating the distance between business processes based on the correlation matrix.On the basis of traditional workflow network,a workflow network with edge weight is defined.Given the rules and algorithms for weighting nodes,we construct the workflow network with weights.Based on the workflow correlation matrix,a workflow correlation matrix with workflow weights is defined,and the definition of the distance between business processes based on the correlation matrix is given.(b)By the comparison of log to log,we give the method of calculating the distance between business processes based on the adjacency matrix.The event log stored in various information management systems is analyzed to be simplified as a simple event log containing only the activity execution trajectory.According to the order of execution of activities,the immediate relations between activities are defined,and then the activity adjacency matrix is constructed.The probability adjacency matrix is defined according to the trajectory incidence and the definition of the distance between business processes based on the activity adjacency matrix is given.