Provenance-based workflow composition with virtual simulation objects technology

Abstract

Accumulated provenance data about correctly constructed and successfully executed scientific workflows provides a valuable source of knowledge, which may be used and reused in new scientific experiments' design. In this paper we propose a provenance-based workflow composition approach - the opposite to package-based one, presented earlier. Processing semantic-compliant provenance data makes available automatic search of already executed workflows (or their parts) in order to reuse them for newly designed applications. This helps the user to be aware only about input and output parameters and variants of intermediate process will be found in provenance and offered for choose. The idea was implemented into Virtual Simulation Objects (VSO) [1] toolbox which provides an intelligent problem solving environment for computational experiments design. Experimental use-case showing the proof of approach hypothesis is also presented.