Validation of Blackboard Systems: On the Order of Knowledge Sources

Daniel O'Leary

There have been few results regarding unique issues of verification and validation of blackboard-based systems. This paper attempts to mitigate this limitation in the literature by focusing on a particular blackboard architecture. In solution-based focusing blackboard systems, problem solving starts by specifying that one of the knowledge sources be instantiated and then iteratively different knowledge sources, typically based on order, are chosen for assistance in developing a solution. Unfortunately, depending on the nature of the particular blackboard system, the specific order in which knowledge sources are chosen can have a major impact on the quality of initial solutions developed. As a result, this paper is concerned with a priori ordering of the knowledge sources so that better initial solutions will be generated. This paper develops some analytical results that can be used to establish partial orderings of the knowledge sources and rules within the knowledge sources, using different relationships between knowledge sources, for economic based systems (i.e., systems where there are costs or revenues associated with the outcomes in the rules).

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.