A logic petri net-based method for repairing process models with concurrent blocks

Abstract

Business processes are usually described and verified by the process models before they are implemented in the information systems. However, there exist deviations between the activities of a process model and those in event logs generated from the information systems. To replay the activities in the logs, the existing model needs to be repaired. In the current repair methods, the models with concurrent blocks cannot be simply repaired. Therefore, this paper proposes a new model repair approach based on logic Petri nets. Concurrent transition pairs and choice transition pairs are constructed based on process trees. By traversing optimal alignments, the transitions between the head and tail of concurrent blocks are determined, deviations are used to judge whether the activities in a process model and event logs have concurrent transition pairs or choice transition pairs, and the model is then repaired via logic Petri nets. Finally, the correctness and effectiveness of the proposed model repair method are illustrated by the experiments.