Abstract:

In the light of the feature of workflow exception and ontology, a method of exception recognition based on Web Ontology Language (OWL) and propertydriven and task property was proposed. This method needs several properties, which were extended by transaction properties of workflow Activity Model (WAMO), being added to one task, and described by the OWL. Then several mapping relations between task exceptions and handling patterns via properties were established. When the task throws some exceptions at runtime, their handling patterns in exception ontology can be found by the OWL reasoner through these mapping relationships that have been defined. Then the handling mechanism takes corresponding measures to resolve these exceptions. Because the definition of OWL is incremental and possesses Uniform Resource Identifier (URI), this method can not only make exception handling mechanism recognize handling patterns of task exception through semantic rather than key words, but also can be adapted to extended development for exception knowledge repository in the distributed environment, which can reduce its development workload.

Key words: artificial intelligence, workflow, exception, Web ontology language(OWL), ontologybased reasoning