An Incremental Approach to Modeling Flexible Workflows Using Activity Decomposition and Gradual Refinement

摘要

For existing workflow models, it is still difficult totake into account both the hierarchical relations amongactivities and their execution orders at the same time, aswell as the dynamic uncertainties of workflow processes.Aiming at these problems, an approach to modeling flexibleworkflow using activity decomposition and incrementalrefinement is presented. First, the activities as well as theirdecomposition relationship and the decomposition rules areanalyzed to establish an activity decomposition model,where flexible activities, with verifying constraints, are usedto package uncertainties. Then temporal relations areintroduced into the model to ensure the proper executionorders of activities, which are tightly integrated with thehierarchical relationship as well as different granularactivities to make the model perform well. The executionmechanism of the model, such as the delivering of eventsand states constraints of parent-child activities, is discussedin detail, and a common algorithm for activity execution isalso presented. Finally, the proposed approach is applied tothe PBL learning system, and the results indicate theeffectiveness of the proposal.