Mining process models from event logs in distributed bioinformatics workflows

Designing a workflow is a complicated process, and there are discrepancies between the actual workflow processes and the ideal processes. So we have researched how to discover workflow models, and we have applied the techniques to distributed bioinformatics workflows. The source for such techniques is a "workflow log" containing information about the workflow process. We presented a new algorithm to extract a process model from such a log and represented it in terms of a Petri net. The algorithm can handle concurrence and recurrence of the distributed bioinformatics workflow processes that are the restrictions of other algorithms. The a algorithm can successfully mine any workflow represented by WF-net.