A high-level Petri nets-based approach to verifying task structures

As knowledge-based system technology gains wider acceptance, there is an increasing need for verifying knowledge-based systems to improve the reliability and quality. Traditionally, attention has been given on verifying knowledge-based systems at the knowledge level, which only addresses structural errors such as redundancy, conflict, and circularity in rule bases. No semantic error such as inconsistency in the requirements specification level has been checked. In this paper, we propose the use of task structures for modeling user requirements and domain knowledge at the requirements specification level, and the use of high-level Petri nets for expressing and verifying the task structure-based specifications. Issues in mapping task structures into high-level Petri nets are identified, for example, the representation of task decomposition, constraints, and state model; the distinction between follow and immediately follow operators; and the composition operator in task structures. The verification of task structures using high-level Petri nets is performed on model specifications of a task through constraints satisfaction and relaxation techniques and on process specifications of the task based on the reachability property and the notion of specificity.