The well-behaved properties in workflow systems are very important for business processes. In this paper, a structured business process modelling approach by using free-choice Petri nets and workflow analysis techniques is proposed. This approach can be used to verify the soundness of workflow process specification for finding control-flow errors in the designs. For example, deadlock and dangling tasks in process models can be avoided by using this approach. However, for arbitrary Petri nets, the checking of sound properties is a NP-hard problem. This paper presents a structured business process modelling algorithm. This algorithm employs six well-behaved control structures, chaining rule and modified nesting rule to construct well-behaved process models. The resulting process models from this algorithm are free-choice workflow-nets, for which sound property can be decided in polynomial time. The proposed method is an extension of the workflow-net. The basic transformation rules in the workflow-net are subsets of the proposed nesting rule in this paper. The process models resulted from this structured business process modelling algorithm also have the advantages of modularity, readability and maintainability.
Workflow modelling and analysis Petri nets
Structured business process models CIMOSA