Comparison

To adapt to the changes occurring at run time, most current WfMSs adopt the version control mechanism. A tiny modification, however, might produce a new version of the workflow, and several workflow instances of various versions might run simultaneously. Stanley’s work [5] provides the adapter mechanism to allow each organization maintaining its services, or, the process definitions it involved. They may replace the version control mechanism with the dynamic binding of the services.

Compared with their models, the advantages of our system are the adaptation capability, where the process agent of each workflow instance can adapt to the changes at run time, and the practicability, where the most business logics can be enacted by our system. Rather than being restricted to the control flow as Aalst’s model [19], we cover the changes both on data flow and control flow. In our system, the user will not be aware of the agent system, and the detection and resolution of the dynamic changes are transparent to the user, similar to Kwak’s framework [8]. As described in chapter 2, the obligation rules of Abrahams’ approach [20] have to be defined carefully to avoid the divergence of the number of the rules. It is not necessary in our system to set the rules to trigger each other rules for the flexibility. By the support of the underlying agent system, we can easily achieve the flexibility by the agent’s reasoning mechanism.

5.2 Conclusion

The more quick Internet technologies evolve, the more complex the Internet

software grows. Because requirements are changed rapidly, most software costs arise from re-developing. To reduce costs and unnecessary reworking, software flexibility and adaptability are now significantly considered. In this paper, we discuss the modifications of a workflow and present an approach to increase adaptability on an agent-based WfMS. An intelligent agent can maintain its knowledge and react to variant environmental changes along the workflow. Adopting software agents helps the WfMS create, execute, and manage workflow processes more flexibly. We have classified all kinds of dynamic changes on process definitions, which are analyzed and categorized according to role, process, and artifact sub-models. Furthermore, solutions to adapting each change type are proposed respectively. An efficient algorithm to differentiate two process definitions is also well-defined. The designed agents therefore can react to these changes at runtime.

Environmental changes include not only those on user requirements, but several kinds of resource exhaustion or unexpected hardware/software failures. Some of them are not caused by humans and beyond our scope in this paper. However, more consideration and designs for handling unexpected events may bring a software system more solid and stable. We'll advance the adaptability of the WfMS and take more factors into account in the future.

References

[1] Gregory Alan Bolcer and Richard N. Taylor. Advanced Workflow Management Technologies. Journal of Software Process Practice and Improvement, 1999.

[2] Marcus J. Huber. JAM：A BDI-theoretic Mobile Agent Architecture.

Proceedings of the third annual conference on Autonomous Agents. Seattle, United states, 1999.

[3] Jorge Cardoso, Zongwei Luo, John Miller, Amit Sheth and Krys Kochut.

Survivability Architecture for Workflow Management Systems. Proceedings of the 39th Annual ACM Southeast Conference, Athens, Georgia, 2001.

[4] Yuhong Yan, Zakaria Maamar, and Weiming Shen. Integration of Workflow and Agent Technology for Business Process Management. Computer Supported Cooperative Work in Design, The Sixth International Conference on, 2001.

[5] Jie Meng, Stanley Y.W. Su, Herman Len and Abdelsalam Helal. Achieving Dynamic Inter-Organizational Workflow Management by integrating Business Processes, Events and Rules. Proc. of the 35^th Hawaii International Conference on System Sciences. 2002

[6] N.R. Jennings, P. Faratin, T.J. Norman, etc. ADEPT: Managing Business Processes Using Intelligent Agent. Proc. of BC Expert Systems 96 Conference. 1996

[7] GMD FOKUS and IBM Corporation. Mobile Agent System Interoperability Facilities Specification. Nov. 1997 Standard proposal.

[8] Myungjae Kwak, Dongsoo Han and Jaeyong Shim. A Framework Supporting Dynamic Workflow Interoperation and Enterprise Application

Integration. Proceedings of the 35^th Hawaii International Conference on System Sciences. 2002

[9] Håvard D. Jørgensen. Interaction as a Framework for Flexible Workflow Modeling. Proceeding of the 2001 International ACM SIGGROUP Conference on Supporting Group Work.

[10] Müller, Robert and Rahm, Erhard. Dealing with Logical Failures for Collaborating Workflows. Etzion, O.; Scheuermann, P. (Eds.): Proceedings CoopIS 2000, Eilat, Israel, September 6-8. LNCS 1901: 210-223.

[11] Workflow Handbook 2003. Edited by Layna Fischer. Published by Future Strategies Inc. with WfMC. 2004. ISBN 0-9703509-4-5

[12] Paul Dourish. Developing a Reflective Model of Collaborative Systems.

ACM Transactions on Computer-Human Interaction. Mar 1995.

[13] Sea Ling and Seng Wai Loke. Advanced Petri Nets for Modeling Mobile Agent Enabled Interorganizational Workflows. 9^th Annual IEEE International Conference and Workshops on the Engineering of Computer Based Systems (ECBS’02).

[14] Gary J. Nutt, Scott Brandt, Adam Griff, and Sam Siewert . Dynamically Negotiated Resource Management for Virtual Environment Applications.

IEEE Transactions on Knowledge and Data Engineering. Jan/Feb 2000.

[15] Ray J.R. Lin. Enacting a Software Development Process. Jun 1996.

[16] Gary J. Nutt. The Evolution toward Flexible Workflow Systems. Distributed Systems Engineering, Dec 1996.

[17] Workflow Management Coalition. Workflow Management Coalition terminology and glossary (WFMC-TC-1011). Technical report. Workflow management coalition. Feb 1999.

[18] Anind K. Dey, Gregory D. Abowd, and Andrew Wood. Cyberdesk: A

Framework for Providing Self-Integrating Context-aware Services. ACM Symposium on User Interface Software and Technology. 1997.

[19] W.M.P. van der Aalst. Generic Workflow Models: How to Handle Dynamic Change and Capture Management Information? International Conference on Cooperative Information Systems. 1999.

[20] Alan Abrahams, David Eyers, and Jean Bacon. An Asynchronous Rule-Based Approach for Business Process Automation Using Obligations. Third ACM SIGPLAN Workshop on Rule-Based Programming. 2002.

[21] Weishuai Yang, Shanping Li and Ming Guo. Mobile agent: Enhancing Workflow Interoperability. International Conference on Info-tech and Info-net. 2001.

[22] Chiung Wen, Chang and Feng Jian, Wang. Using Software Agents to Design a Modern Workflow Management System. Master thesis. Dept. of CSIE, National Chiao Tung University. 2004.