With the ever-increasing process automation and the growing interconnectivity among the systems – especially with the advances of the Internet - the need for computers to support the automatic matching of workflows is greater than ever. One of the obstacles for the automatic workflow matching is the problem of semantic matching (i.e. identifying concepts with similar meanings but expressed in different forms).
The WoSeMaP project deals with the problem of semantic matching of workflow definitions; two or more workflows systems defined in different representations need to interact (e.g. reuse, integrate, interoperate, etc.) amongst them. Nowadays, the solution to this problem is provided by a human expert team that maps the concepts among the representations. This process is not just time and effort consuming, but error prone and hard to maintain.
This project aims to produce a Workflow Semantic Matching Platform (WoSeMaP) to support the semi-automatic finding of semantic similar workflows.
The WoSeMaP focused in two challenging problems of the workflow semantic matching:
- The ontological differences of the workflows
Two workflows modeled by different entities (i.e. persons, organizations, enterprises, etc.) normally include the same terms (i.e. names, descriptions, identifiers) with different meaning and/or use different terms to express the same concept.
- The second problem engaged by WoSeMaP project is the granularity differences between the workflows.
Two workflows expressed by different entities can have different granularity level of process descriptions, meaning that a task in one workflow could be represented by many tasks in another workflow.
The WoSeMaP goal is to research and create the theoretical and practical tools to support workflow semantic matching, by dealing with these two challenges. Currently, the developed tools are:
- Semantic Similarity Networks (SSN)
The SSNs are designed to represent the information contained in the workflow’s ontologies and calculate the semantic similarity among all concepts of the two workflows.
- The WoSeMaP Workflow Similarity Metrics
The WoSeMaP is focused on producing a set of workflow similarity metrics that allow identifying the similar sequences of tasks with similar context and results. The goal is to compare groups of tasks instead of one task at a time, reducing the granularity challenge.
Combining the SSNs and the workflows similarity metrics, the WoSeMaP would be able to provide semi-automatic support for the solution of the Workflow Semantic Matching problem.
Two members of the Wi2-group are working within WoSeMaP project.
- Prof. Dr. Ralph Bergmann
- M.Sc. Fawsy Bendeck
This project has received financial support from
- Der Deutsche Akademische Austauschdienst (DAAD) e.V.
- The University of Trier
- Fawsy Bendeck: Three Fold "Ontologie + Model + Instance (OMI)" Semantic Unification Process. In International Conference on Advances in Internet, Processing, System and Interdisciplinary Research (IPSI-2004) , Stockholm, Sep. 2004, ISBN 86-7466-1173.