Conformance checking is able to detect deviations in business process execution. An online detection capability is required to anticipate and respond immediately to possible impacts. The state-of-the-art online conformance checking is a prefix-alignment (PA) technique. However, this technique has an important weakness of maintaining all the state data of running processes in memory. In an online environment, the last event of a case is unknown, whereas a PA requires this information to free up memory space for other cases. Consequently, the PA does not meet the requirements of online conformance checking to process infinite data (event stream) without memory constraints. PA also has a conplex state space search computation especially for large and complex reference process model. In this paper, a Graph-Based Online Token Replay (GO-TR) method is proposed. This method takes benefit of Graph Database to adapts the Token-Based Replay (TBR) technique which has simple replay computation. We propose a Replay Image (RI) to store the case administration and developed a cypher based algorithm to simulate token replay on the RI to handle the event stream. Finally, we propose a cypher-based algorithm to identify and replay invisible path. The experiment result show that GO-TR has succeeded in adapting TBR and solved the problem of wrong-place tokens in TBR. GO-TR outperform the replay throughput for relatively low amount of data against state of the art online conformance checking. In terms of memory usage, GO-TR shows its superiority over state of the art because it is safe against memory limitations problems.
Read full abstract