An effective architecture for Decision Support Systems

Abstract

Decision Support Systems (DSS) are, by nature, general- purpose systems, because they must support a variety of managers who have different decision styles and different problems. However, it seems that no effective general-purpose DSS have yet come into existence, although the components of DSS such as data base technology, modeling techniques, inexpensive graphic display etc., have progressed to the point where we should now be able to build effective DSS. This shortcoming seems to result from the following fact: Research on decision support has focused on data enlargement and model refinement, however, little attention has been paid to DSS architecture which integrates these components of DSS. It has not been well appreciated that DSS architecture itself facilitates learning about unstructured-problem solving and enables system evolution. In this paper, we propose a DSS architecture based on the study of unstructured-problem solving and considerations of the needs of managers as non-computer specialists. We illustrate this with a system realized using this architecture.