Module base design for a portfolio of institutional decision support systems

Recently, there has been a trend toward utilizing artificial intelligence techniques in decision support systems (DSSs) to enhance system capabilities and support. This paper focuses on using one such technique (productions or rules) for enhancing module representation, access, execution and maintenance flexibility in a class of systems called institutional DSSs. In the approach we propose, an organization maintains a portfolio of such DSSs. An individual system contains, in its knowledge base, only the "structural", "invocational" and "presentation" knowledge of the modules used by that system. Executable representations of all of the modules in the portfolio and the associated "procedural" knowledge are generated and stored in an external organizational model base (OMB). We discuss representation schemes for the three types of module-related knowledge in each DSS as well as the architecture for the system-user interface.