Reducing Anomaly Detection in a TS Workflow

Abstract

Nowadays workflow management systems (WfMS) are widely used for improving business processes and providing better quality of services. The analysis of workflow can facilitate locating problems in business processes and prevent repeated errors during workflow execution. A temporal workflow is described with the min and max execution time intervals for each process and modeled by control structures (Sequence, AND, XOR, Loop). These time intervals are essential factors for the analysis and anomaly detection in a workflow. In our previous work, we provided redefined definitions for anomalous behaviors and algorithms to detect anomalies within a temporal workflow. However, an essential factor, a loop structure and its temporal effects are not discussed yet. In this paper, our work extends the analysis to the workflows containing loop structure. To simplify the anomaly detection inside the loop, we first transform the loop into an XOR structure branch, perform a series of analyses and then design an algorithm to find the anomalies more efficiently. We also make an analysis for the contribution due to the algorithms.