Download presentation
Presentation is loading. Please wait.
Published byMartha Tate Modified over 9 years ago
1
Dr Marcello La Rosa BPM Research Group, Queensland University of Technology
2
Fact Enterprises in various industries tend to collect thousands of business process models over time: Thousands of activities and related artefacts Numerous stakeholders with different skills and responsibilities Difficult to keep track of such volumes of models (“model management”), especially due to: Overlapping content across models Evolving content over time (legacy models vs. new versions) Different modelling notations, e.g. EPCs, BPMN, Protos, BPEL... Different modelling purposes and granularity Different modelling guidelines
3
Opportunity Research in BPM has flourished over the past decade: 1. Algorithms to evaluate the quality of models (e.g. correctness, structuredness) 2. Experiments to establish the factors that contribute to the understanding of process models by users 3. Algorithms to identify similarities between process models 4. Techniques to customize process models (e.g. configuration, adaptation) However, these techniques and tools look at process models in isolation, rather than viewing a process model in relation to other process models, or deal with small collections of models.
4
Proposal: AProMoRe Build an open and extensible platform for the storage and disclosure of process models which provides advanced features to deal with large process models and collections thereof. Joint collaboration among 6 universities: QUT TU/e University of Tartu Humboldt University of Berlin University of Grenoble Hasso-Plattner Institute
5
Service Areas Evaluation establish adherence to various quality notions (e.g. correctness criteria, benchmarking frameworks) Comparison search for similarities (e.g. conformance to reference models/patterns, track extensions) Management control the creation and evolution of process models (e.g. via configuration, merging, improvement) Presentation improve the understanding of process models (e.g. contextualization via abstraction and colouring mechanisms)
6
Canonical format: the power of losing Common process representation as directed attributed graph: 1. Standardization – cross-language operations can be performed directly and concatenated 2. Interchangeability – non structural aspects captured by meta-data swap notations or semantics at (almost) no cost 3. Flexibility – inheritance mechanism different algorithms can work at different abstraction levels
7
Canonical format
8
Conversion chart (fragment)
9
Canonization – example XOR split AND split XOR join XOR split
10
Architecture
11
Data layer Services: Portal, Manager, (De)Canonizer, Data Access XML Schema for Canonical Process Format and Annotation Facade classes to manipulate CPF, ANF, YAWL, EPML, XPDL... Facade-based (de)canonizer: Full EPML 2.0 (EPCs, eEPCs, iEPCs, C-EPCs, C-iEPCs) Full XPDL 2.1 (BPMN 1.0, 1.1, 1.2) WS functionalities (import, export, simple filter, edit) Oryx integration (XPDL 2.1) Gather model collections Current implementation
12
Model collections YAWL Order fulfillment (C) * Production * Municipalities (C) * EPC Video post-production (C) * Land development (C) * Suncorp insurance SAP R/3 iEPC Audio post-production (C) * BPMN Land development (C) * Traumatology Airports (C) * Protos Municipalities (C) *
13
Can I help? Yes you can! Available projects: Integration with ProM: import/export + log storage SVN Support Security layer (authentication, authorization, non disclosure) Support for other languages (WF-Nets, YAWL, Protos, BPEL...) Build your algorithms in it! Donate your content! (logs, models...) © 2009, Marcello La Rosa
14
Marcello La Rosa Business Process Management Group Faculty of Science & Technology Queensland University of Technology 126 Margaret Street Brisbane QLD 4000 Australia p 07 3138-9482 e m.larosa@qut.edu.au w www.marcellolarosa.com
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.