Detection of semantically equivalent fragments for business process model change management

Modern business process modeling environments support distributed development by means of model version control, i. e. comparison and merging of two different model versions. This is a challenging task since most modeling languages support an almost arbitrary creation of process models. Thus, in multi-developer environments, process models or parts of them are often syntactically very different but semantically equivalent. Hence, the comparison of business process models must be performed on a semantic level rather then on a syntactic level. For the domain of business process modeling, this problem is yet unsolved. This paper describes an approach that allows the semantic comparison of different business process models using a normal form. For that purpose, the process models are fully automatically translated into process model terms and normalized using a term rewriting system. The resulting normal forms can be efficiently compared and easily be used for reconciliation. Our approach enables the semantic comparison of business process models ignoring syntactic redundancies. © 2010 IEEE.