The Study Of Detecting And Repairing Methods For Business Process Systems Based On Petri Nets

Successful business process system management depends on effective model and analysis.How to guarantee the behavioral consistency between different business process models has become a bone of contention in existing studies.However,these studies cannot help us calculate the behavioral consistency accurately when there are special modules and data information in business process models.Besides,the behavioral consistency of two models may be affected by guard functions,data dependencies and different kinds of data-flow errors on transition sets.Given a model without errors,some errors can still occur during the process execution due to specific system requirements.Therefore,measuring the behavioral consistency effectively between different models,detecting and repairing data-flow errors as well as satisfying complex system requirements have become three key problems in business process system management.To solve these problems,we analyze the model of workflow net(WF-net)and workflow net with data(WFD-net)based on the basic theory of Petri net.Firstly,we measure the behavioral consistency between two WF-net models with special modules.Secondly,we study the behavioral consistency between two WFD-net models with data operations and guard functions.After then,detecting and repairing all kinds of data-flow errors in WFD-net models are provided.At last,we provide the design by constructing missing data based on the system requirements(i.e.,control the firing times of critical transitions).The main contents and contributions of this paper are given as follows.(1)For the WF-net models with special modules,the existing methods cannot calculate the behavioral consistency degree of them accurately.Therefore,this paper provides new consistency measurement method.Based on the existing studies,this method refines three kinds of correspondence relations in WF-net models.Then,their behavioral relation matrices are constructed.Based on them,BCDC method is proposed to measure the behavioral consistency degree of two WF-net models with/without special modules from four angles.A group of experiments is done to show that our consistency measuring methods are more accurate than the existing methods once there are special modules.(2)For the WFD-net models with guard functions and data dependencies,the existing studies cannot measure the behavioral consistency degree of them accurately.Therefore,we provide the methods of constructing behavioral relation matrices of WFD-net models.As the behavioral relations of transition pairs would be affected by data dependencies,we redefine the behavioral profile based on the original behavioral relations of transition pairs.The behavioral relations of transition pairs in a WFD-net model are uncertain when the values of guard functions are different.Therefore,we point out that the consistency degree of two WFD-net models is not unique.We solve this problem by establishing multiple behavioral relation matrices of WFD-net models.A group of experiments is done to show that our consistency measuring methods are more effective.(3)For the behavioral consistency measurement of two WFD-net models is affected by data-flow errors,the existing methods cannot repair errors accurately.Therefore,this paper provides the methods of detecting and repairing data-flow errors in WFD-net models.Firstly,the formal definitions of four kinds of data-flow errors(i.e.,redundant data,missing data,lost data and inconsistent data)are given based on order relations.After then,we reveal the relations between these errors and organize them into a hierarchy.At last,we propose some new methods to repair data-flow errors in a WFD-net model based on the level of the hierarchy,system requirements and repair strategies.A group of experiments is done to show that our methods can repair data-flow errors in WFD-net models with delete operations and weak circulation relations effectively.(4)To satisfy the system requirements of a WFD-net system,we provide methods named construct missing data.First of all,this method refines the missing data and defines two kinds,i.e.,direct missing data and indirect missing data.Then,we search for the critical transitions based on the system requirements,and these two kinds of missing data are constructed.Finally,a corresponding algorithm is proposed to minimize the scale of missing data.The methods in this paper show that the design requirement of our systems can be satisfied effectively based on the analysis of real cases.