|
|
Events such as the "AI meets the Real World'' Workshop recently held in Newark, New Jersey have shown that a number of AI technologies have reached a level of maturity which sees them being used to solve a number of real world tasks in companies such as Boeing, NASA, Lucent and Lockheed Martin. However, while there have been a number of notable successes the impact of AI technologies has been less than expected. One area which seems ripe for exploitation is process automation or workflow which is at the heart of many organizations and businesses. Workflow Management Systems are integrated software tools for supporting the modeling, analysis, and enactment of business process. The technology's development has been driven by the move to process oriented management in the 90s via initiatives like "Continuous Business Process Improvement'' and "Business Process Re-engineering''. The market for workflow management software had grown from around $100 million in 1991 to $2.5 billion in 1996. From a research perspective DARPA has identified workflow as one of its key "must have" technologies and is investing heavily in developing the next generation workflow systems for the military. However, the impact of such systems will go well beyond the military and will be of great interest to the general business community. The Object Management Group has recently established an industry standard for interoperability of workflow systems, opening the door to enterprise-level and inter-enterprise process automation. Such standards will promote the development of specialized workflow systems and components incorporating AI functionality.
Conventional workflow management systems use explicit models and representations of process, along with automated tools that support the activation and ongoing management of a process instance. This technology has to date found application only in areas characterized by simple administrative type processes such as insurance claim processing. The benefits alluded to by workflow technology are highly desirable and the workflow research community has set the agenda of developing techniques that enable these benefits to be achieved in applications characterized by complex tasks performed in dynamic and uncertain environments. these are precisely the classes of tasks and environments that AI research has been investigating in the context of controlling computational entities and physical devices.
While workflow has emerged over the last eight years the artificial intelligence (AI) community has been involved with related research on process management for several decades. In contrast to workflow's focus on business and manufacturing processes, the AI community has been motivated primarily by domains that involve reactive control of computational entities and physical devices (e.g., robots, antennas, satellites, computer networks, agent communities). Despite these differing concerns and perspectives, there is much overlap between the objectives, requirements, and approaches of these two communities. Workflow provides the business drivers and the computational infrastructure that respectively motivate and enable an industrial deployment of AI technology. Such application will further develop the commercial credibility of AI technology while simultaneously providing immediate feedback to shape research programs.
The objective of this Workshop, is to bring together researchers, practitioners, and applied AI specialists from diverse fields to discuss issues and emerging technologies for developing workflow and process management systems. In particular, we hope the workshop will address innovative ways of putting intelligence into Workflow Management to Revolutionize Business.
Day 1 Sunday August 1, 1999 (8:30 - 18:00)
08:30 - 09:00 Welcome and Introduction, Mamdouh Ibrahim.
Session 1: Framework and Architecture: 09:00 - 10:00
Moderator: Fred Cummins
- A Planning Framework for Workflow Management, Clemens Beckstein and Joachim Klausner
- A Scalable Heterogeneous Architecture for Agent- Oriented Workflow Management, John Hickie, Broadcom Eireann Research
Coffee: 10:00 - 10:30
Session 2: Infrastructure: 10:30 - 11:30
Moderator: Mamdouh Ibrahim
- Building a Peer-to-Peer workflow infrastructure, Paul Santanu and David Hutches, IBM
- A Generic Approach to Component Based Workflow Control, Christian Mittasch and Alexander Schill, Institut fur Informatik, Germany
Session 3: Workflow and Process Management in Business: 11:30 - 13:00
Moderator: Fred Cummins
- The NetAcademy - Managing Internet Peer Review Process In a Multi-agent Framework, Lei Yu, University of St. Gallen, Switzerland
- Knowledge Based Modelling of Processes, Almundenasierra-Alonso, Ricardo Aler and Daniel Borrajo, Universidad Carlos III de Madrid
- Business Process Modelling and Design. AI Models and Methodology, Manolis Koubarakis and Dimitrios Plexousakis, University of Crete
Lunch: 13:30 - 14:00
Session 4: Process Management: 14:00 - 15:30
Moderator: Brian Drabble
- KPM: Knowledge Based Process Management,: Susan Lander, Knowledge Technologies International
- Agent-Based Process Management, Petri, Goldman, Raquel
- A Business Process Agent, John Debenham University of Technology, Sydney
Coffee: 15:30 - 16:00
Demonstration Session (Open for all workshop participants): 16:00 - 16:40
- Demonstration, Jan Erik Ressem, Computas AS, Norway.
The PLANET Initiative : 16:40 - 17:00
- Paul Kernay, BT.
Invited Speaker: 17:00 - 18:00
- "Workflow in a Community Knowledge System", Danny Bobrow.
Day 2 Monday August 2, 1999 (8:30 AM - 5:15 PM)
Invited Paper: 08:30 - 09:00
- Adaptive Workflow: On the Interplay between Flexibility and Support, Wil van der Aalst, Twan Basten, Eric Verbeek, Peter Verkoulen, and Marc Voorhoeve Eindhoven University of Technology.
Invited Speaker: 09:00 - 10:00
- Martha Pollack, University of Pittsburgh, USA.
Coffee: 10:00 - 10:30
Session 5: Reactive and Proactive Control: 10:30 - 11:30
Moderator: Brian Drabble
- SWIM: An AI Based System for Workflow Enabled Reactive Control, Pauline Berry, SRI International and Brian Drabble, CIRL.
- Distributed Workflow Coordination by Proactive Software Agents, B.R. Odgers, J.W. Shepardson and S.G. Thompson, BT Laboratories, UK.
Session 6: Multi-Agents: 11:30 - 13:00
Moderator: Brian Drabble
- Multiagent Workflow Management, Feng Wan, Sudhir K. Rustogi, Jie Xing, and Munindar P. Singh, North Carolina State University
- Multi-Agent Framework for Facilitating Information Extraction, C. Curtis Crtmill and Robin Cohen, University of Waterloo
- Meta-Agent: A Workflow Mediator in Multi-Agent Systems, Michal Pechoucek, Czech Technical University
Lunch: 13:00 - 14:00
Session 7: Process Modeling and Design: 14:00 - 15:30
Moderator: Fred Cummins
Coffee: 15:30 - 16:00
Session 8: Learning and Supporting Techniques: 16:00 - 17:00
Moderator: Mamdouh Ibrahim
- An Inductive Approach to the Acquisition and Adaption of Workflow Models, Joachim Herbst, Daimler Chrysler AG, Germany
- Towards Learning Notification Triggers, Steffen Staab & Thomas Barnekow, AIFB, University of Karlsruhe
- Applying Local Search to Structural Constraint Satisfaction, Alexander Nareyek, Research Institute for Computer Architecture and Softwaretechnology, Germany
Conclusion and Closing Remarks: 17:00 - 17:15: Brian Drabble
This workshop will be held as part of the Sixteenth International Join Conference Artificial Intelligence (IJCAI-99) in Stockholm, Sweden during late July and early August 1999. Please see the conference site for more information.
Page maintained by peter.jarvis@ed.ac.uk, Last updated: 25 June 1999.