Intelligent Agent-Based Flexible Workflow Management for Crisis Resolving

The unpredictability of business processes require workflow systems to support crisis resolving in the dynamical environment. Traditional approaches to handle this problem have fallen short, providing little support for changes, particularly when the execution of a process has started. Further more, crises vary widely in their characteristics and significance, challenging the application of any single approach to resolving them. In this paper, we briefly discuss the classification of such crises, highlighting differing impacts on the workflow model. Based on this discussion, we define our main goals to address in the development of flexible workflow support, including strategies for avoiding crisis, detecting them when they occur, and resolving them at various levels of impact. We then proposed a novel intelligent agent-based architecture to supporting these goals within the design of a workflow system infrastructure, by utilizing crisis manger agents, crisis monitor agents, crisis backup agents and crisis mender agents, and integration with an intelligent agent-assisted decision support system.