Research On Recommendation And Optimal Alignment Of Business Process Modeling Based On Petri Net

With the development of modeling ideas in enterprises,business process modeling has reached a very complex level.In order to assist modelers to quickly build higher quality business process models,the research on process modeling recommendation techniques has arisen.Most of the existing methods are based on graph structure or process behavior.However,there may be various graph structures and different paths to achieve the same business target in the actual work-flow,so the existing recommendation technology may lead to the loss of a lot of useful recommendation information.Therefore,some scholars proposed that considering the process target in the modeling process can provide strong guidance for modelers,but so far,there are few recommended modeling methods that relate the process target.At the same time,with the development of information technology,today’s information systems can accurately record the execution of business processes.Therefore,when the new business process model is put into use,it is necessary to analyze the deviation between the actual process and the modeling process,that is,compliance verification,and then implement process enhancement,performance analysis,log repair and other operations on this basis.At present,the alignment of process model and event log has become the accepted standard to solve the compliance verification problem.The optimal alignment refers to the alignment with the lowest cost for a given cost function.However,the existing cost function may lead to the problem of alignment cost seriously deviating from perceived cost.In addition,relying only on the cost function to calculate the optimal alignment between the model and the log may cause the alignment results to fail to accurately reflect the deviation between the actual process and the modeled process.Firstly,based on the concept of Petri net and its behavior profile,this paper proposes a new process modeling recommendation technique,and reviews the behavior relationship of the new model.Then,when the modeling process is aligned with the actual process,a new cost function and a new optimal alignment selection strategy are proposed.The main research contents are as follows:(1)Aiming at the limitation of graph structure and behavior path based recommendation technology,a process modeling recommendation method based on target rules defined by behavior profile was proposed from a new perspective.This method is guided by the business target expressed by the behavior profile,aiming at extracting more useful recommendation information from the process repository for modelers,and then assisting modelers to build higher quality business process models.(2)Aiming at the problem that new business process model may have behavior relationship error,a method to review new business process model from global perspective is proposed.In this method,the concept of behavior profile is used to review the behavior relationships among activities in the new model based on process repository,so as to ensure the correctness of the behavior relationships among activities in the process.The results of the review can be used to optimize the new model,and then get a new business process model with higher quality.(3)To solve the problem that the existing cost function may not accurately reflect the deviation between the actual process and the modeling process,an effective alignment method between the process model and the event log based on perceived cost was proposed.Based on the concept of perceived cost,a new cost function,namely the important synchronization cost function,was defined.Under this function,the final alignment cost is closer to the perceived cost and the alignment result is closer to the proposed execution track of the non-conformance log trace,so as to more accurately reflect the deviation between the actual process and the modeling process.At the same time,an efficient alignment method is proposed based on this function.(4)Aiming at the limitation of relying only on cost function to calculate model and log optimal alignment,an optimal alignment selection strategy of business process model based on log probability distribution was proposed.In this strategy,optimal alignment is no longer only the alignment with the lowest cost under a given cost function,but combines the alignment cost with the probability distribution of the conformance log to reflect the confidence that an alignment is optimal,and takes the alignment with the highest confidence value as the optimal alignment.This optimal alignment selection strategy can solve the problem of unreasonable alignment results and the optimal alignment is not unique,so that the alignment results more accurately reflect the deviation between the actual process and the modeling process.Figure [31] Table [15] Reference [86]...