Process Synthesizing for Benchmarking - A Macro Analysis of Process Models

Thursday, January 15, 2015 - 10:30
University of Lugano, room SI-015, Informatics building (Via G. Buffi 13)

Workload defines the benchmark to a great extent. It needs to be realistic and well designed in order to address the system’s main features. In the case of benchmarking Workflow Management Systems the workload needs to address numerous different system features.

Therefore, we study the automated process synthesis in order to address different user-defined characteristics and make the workload sets useful for different use-cases. With respect to this goal we need to study large process models collections for similarities. Currently, there are three main research streams on process models similarities: text similarities, behavioral similarities and structural similarities. As behavioral data and text semantics are usually missing from mock-up or obfuscated models we focus on the detection of structural similarities.

This information gives us a macroscopic overview of the existing collection. The problem of the automated detection of re-occurring structures in a collection of process models, when text semantics or behavioral data are missing is a case of (sub)graph isomorphism, which is mentioned as NP-complete in the literature. In this seminar we will focus on the methodology followed for the detection of re-occurring structures in a collection of process models. We will also demonstrate the first results from the method’s application.

Marigianna Skouradaki is a Research Associate and PhD Candidate at Institute of Architecture of Application Systems at University of Stuttgart (Germany) under the supervision of Prof. Frank Leymann. Before that she received an M.A. form University of Leeds (UK) in 2010, M.Sc. and B.Sc. from University of Crete (Greece) in 2012 and 2009, respectively. She is currently involved in BenchFlow project, which targets to the creation of a standard benchmark for BPMN 2.0 Workflow Engines. Her research interests focus on the creation of realistic workload models that satisfy different benchmark scenarios. For this she studies process similarities, process synthesizing, process re-usability, workflow engines benchmarking and workload generation. Other research interests include Information Communication Technologies and adaptive learning.