|Ph.D Student||Lincoln Maya|
|Subject||Framework and Tools for the Generation and Customization of|
Business Process Model Content
|Department||Department of Industrial Engineering and Management||Supervisors||Professor Avigdor Gal|
|Dr. Mati Golani|
|Full Thesis text|
In recent years, researchers have become increasingly interested in developing frameworks and tools for the generation and customization of business process model content. This thesis proposes a framework that uses an existing process repository as a reference model against which typical Business Process Management (BPM) activities can be performed effectively. In particular, we suggest three methods within this scope aiming at improving the usage of business process repositories. The first method focuses on a machine-assisted design of new process models, based on business logic that is extracted from real-life process repositories using a linguistic analysis of the relationships between construct of process descriptors. The suggested method can assist process analysts in designing new business processes while making use of knowledge that is encoded in the design of existing process repositories. The second method focuses on content-based validation of business process models, based on existing organizational policies. This methodology goes beyond structural notation and proposes to automatically extract business logic from process repositories as a basis for content validation. The thesis proposes a stepwise method for content-based validation that includes deficiency identification (using existing descriptors as a reference), validation score calculation, and generation of a ranked list of corrections. The third method focuses on searching the content of business process models based on the operational meaning of the searched query. The suggested method supports users that opt to explore specific know-how encapsulated in a process model by retrieving suggestions for process model segments that describe how to fulfill the requested business goal based on the an input request phrased in a natural language.