Aligning event logs and process models based on Petri nets

To improve the efficiency of alignment in the process mining, a new alignment approach named RapidAlign was presented between event logs and process models based on Petri nets. The events in the log were observed and the transitions in the model were firing, and the activities in the log and in the model were compared to obtain the log movement, model movement and synchronous movement; the cost value was calculated, and the current states of log and model were recorded; the states with the minimum cost were selected until both log and model arrived at the final states, thus an optimal alignment graph was finally generated. In the graph, the paths from the source node to the target node included all of the optimal alignments between event log and business process model based on standard likelihood cost function. A specific and rigorous characterization was given to illustrate the availability of RapidAlign approach, and its correctness and effectiveness were proved theoretically. After a series of the simulation experiments, the superiority of RapidAlign method was verified. © 2019, Editorial Department of CIMS. All right reserved.