DiscopFlow: A new Tool for Discovering Organizational Structures and Interaction Protocols in WorkFlow

TL;DR

DiscopFlow is a novel tool that enhances Workflow Mining by enabling the discovery of organizational structures and interaction protocols using an extended log meta-model and multi-agent system concepts.

Contribution

The paper introduces DiscopFlow, a new Workflow Mining tool that incorporates organizational perspectives and interaction protocols through an extended log meta-model.

Findings

DiscopFlow successfully discovers organizational structures.

It supports analysis of interaction protocols.

The extended meta-model improves workflow understanding.

Abstract

This work deals with Workflow Mining (WM) a very active and promising research area. First, in this paper we give a critical and comparative study of three representative WM systems of this area: the ProM, InWolve and WorkflowMiner systems. The comparison is made according to quality criteria that we have defined such as the capacity to filter and convert a Workflow log, the capacity to discover workflow perspectives and the capacity to support Multi-Analysis of processes. The major drawback of these systems is the non possibility to deal with organizational perspective discovering issue. We mean by organizational perspective, the organizational structures (federation, coalition, market or hierarchy) and interaction protocols (contract net, auction or vote). This paper defends the idea that organizational dimension in Multi-Agent System is an appropriate approach to support the discovering of this organizational perspective. Second, the paper proposes a Workflow log meta-model which extends the classical one by considering the interactions among actors thanks to the FIPA-ACL Performatives. Third, it describes in details our DiscopFlow tool which validates our contribution.