Reifying Behavior in Decision Support Systems

In order to realize the kind of active, intelligent decision support system advocated by previous authors, we argue that the implementation should be based on reified models of the system behavior, an approach called meta-level reification. In this article, we describe the nature and benefits of meta-level reification for arbitrary computation systems, and relate the benefits to decision support systems. To further illustrate our approach, we present an implementation of a reified architecture based on the MAX architecture (Kuokka, 1990), and show its application to several simple decision support functionalities in an investment domain. Finally, we discuss the advantages, disadvantages and future work of our approach. © 1996 Wiley Periodicals, Inc.