Ontology-Based Exception Handling for Semantic Business Process Execution

摘要

Along with the widespread acceptance of business process management (BPM) and Semantic Web services composition technologies, Semantic Web service oriented programming is becoming an efficient way to develop modern business applications. As Semantic Web services are inherently unreliable, how to develop reliable service oriented applications is a significant and challenging problem, especially in complex, untamed and dynamic services environment. However, current business process programming languages for Semantic Web services provide almost no support for exception handling, and runtime environments are weak in providing reliability and adaptability. In this paper, we propose an ontology based exception handling method for semantic business processes using Semantic Programming Language (SPL). We identity a new exception taxonomy described as exception ontology and devise an event-driven exception events detection mechanism. We also devise an exception handling framework based on Multi agent system for SPL program. Especially, the framework provides forward recovery support with dynamically substituting failed services when an exception arises during execution by semantically equivalent or semantically similar Web services.