A Deadlock Detection Method for Inter-organizational Business Process Based on Role Network Model

Abstract

In the context of inter-organizational business collaboration, structural deadlock detection is an important means of guaranteeing the correctness of an inter-organizational business process model. The existing deadlock detection methods have limitations in handling some aspects of the dynamics and uncertainty arising from inter-organizational nature of the processes. Role Network Model (RNM) was proposed to model knowledge intensive systems and can better represent the dynamic aspects of business processes using roles. But RNM is weak in the explicit description of business processes. Therefore, an extended Directed Acyclic Graph is proposed to better describe the process side complementing RNM. Second, some extended reduction rules targeting process dynamics are proposed to reduce the complexity. Third, a deadlock detection method is proposed to support highly dynamic processes. At both simulation phase and/or run time, deadlock detection and reduction can be performed step by step. The proposed method can not only improve the efficiency of deadlock detection but also find some deadlocks before all the uncertainties are determined. Finally, through a case study, the result demonstrates that this method can effectively achieve deadlock detection for inter-organizational business processes including dynamic and uncertain business processes.