Enabling effective workflow model reuse: A data-centric approach

- We propose a formal data-driven approach (DWMR) to facilitate efficient workflow model reuse.
- DWMR compliments existing control-flow-focused workflow modeling approaches by explicitly storing workflow data information.
- DWMR provides data-driven workflow model search and composition algorithms to automatically satisfy user query requirements.
- The result generated by our approach has a high accuracy in terms of precision and recall, comparing with human experts.

With increasingly widespread adoption of workflow technology as a standard solution to business process management, a large number of workflow models have been put in use in companies in the era of electronic commerce. These workflow models form a valuable resource for workflow domain knowledge, which should be reused to support workflow model design. However, current workflow modeling approaches do not facilitate workflow model reuse as a fundamental requirement, leading to a research gap in effective workflow model reuse. In this paper, we propose a novel approach called Data-centric Workflow Model Reuse framework (DWMR) to provide a solution to workflow model reuse. DWMR compliments existing control-flow-focused workflow modeling approaches by explicitly storing workflow data information, such as data dependency, data task relationships, and data similarity scores. DWMR also provides data-driven workflow model search and composition algorithms to satisfy user query requirements by automatically combining multiple workflow models. We demonstrate the feasibility of the DWMR approach by applying it to data from a well-known industry workflow model repository.