Efficient behavioral-difference detection between business process models

Abstract

Recently, business process management plays a more and more important role on the management of organizations. Business process models are used to enhance the efficiency of management and gain the control of (co)operations both within an organization and between business partners. Organizations, especially big enterprises, maintain thousands of process models and process models of the same category are with high similarity. For example, China Mobile Communication Corporation (CMCC) have more than 8000 processes in their office automation systems and the processes of the same business in different subsidiaries are with some variations. Therefore, to analyze and manage these similar process models, techniques are required to detect the differences between them. This paper focuses on detecting behavioral differences between process models. The current technique takes hours to compare the behaviors of two models in worst cases. Therefore, in this paper, an efficient technique is proposed by comparing dependency sets and trace sets of features in process models. Experiments show that the technique can detect behavioral differences between two models within seconds in worst cases, at least four orders of magnitude faster than existing techniques.