CBR Based Workflow Composition Assistant

Composing a scientific workflow from scratch may be time-consuming, even if the scientist is fully aware of the semantics, the inputs, and the outputs of the expected workflow. Reusing existing services and parts from already composed workflows can aid in reducing the total workflow composition time. However, matching the semantics and the inputs and outputs of these reusable components manually is not an easy task, especially when there are hundreds of such components available. Even components are annotated with information on the semantics of their inputs and outputs, the complex nature of the semantic languages may make manual component selection even harder. In this paper, we propose a Case-Based Reasoning (CBR) approach to assist composition of workflows based on the characteristics of the inputs and the outputs of the reusable workflow components, facilitating user exploitation of existing services and workflows during workflow composition. The architecture can also be extended to utilize the semantics of the various components improving the precision of the identified reusable components.