A Deadlock Detection Method For Cross-organizational Business Process

The business collaborations of organizations are much more frequent and close than before. There is considerable number of dynamic business processes in business process due to the participating organizations’distribution, local autonomy and the openness, the dynamic and uncertainty characteristics of cross-organizational collaborative relationships. A reasonable and correct business process model is important for collaboration between the organizations smoothly.Business process modeling is very complex and error-prone. If the error of the process model can not be found early, it may lead to significant losses even need to pay great repair costs. Therefore, the effective validation of the process model is very necessary and deadlock detection is a quite important aspect for the validation. Although, there have been research works on deadlock detection methods for business processes and have got considerable achievement, the deadlock detection methods do not consider the deadlock detection under the context of the dynamic changes of the process structure at run time.In response to these problems above, combined with to the acyclic graph(DAG) and Pi calculus, a deadlock detection method oriented cross-organizational business process is proposed based on Role Network Model and the process reduction rules for deadlock detection. The main contents in this paper are as follows:(1) On the basis of the Role Network Model and directed acyclic graph (DAG), the extended directed acyclic graph (eDAG) is proposed to better describe the process side for cross-organizational business processes, especially the dynamic process structure.(2) Considering the dynamic processes, the extended reduction rules are proposed supporting the dynamic processes on the basis of the part existing process reduction rules.(3) Based on the researches above and combined with the Mobility Workbench, a deadlock detection method for the cross-organizational business process is proposed to achieve the deadlock detection for dynamic process.(4) The deadlock detection method is validated through three case studies in E-government. E-commerce, as well as the emergency management respectively.The related researches supplement and improve the existing deadlock detection methods in theory, and the researches have broad application prospects in the field of e-govemment, e-commerce, and the emergency management and other cross-organizational business process management aspects.