Knowledge-Based Dialogue in Intelligent Decision Support Systems: Fid-Bau Portal

Knowledge-Based Dialogue in Intelligent Decision Support Systems

Hollnagel, Erik

Abstract The overall goal for the design of Intelligent Decision Support Systems (IDSS) is to enhance understanding of the process under all operating conditions. For an IDSS to be effective, it must: (1) select or generate the right information, (2) produce reliable and consistent information, (3) allow flexible and effective operator interaction, (4) relate information presentation to current plant status and problems, and (5) make the presentation at the right time. Several ongoing R&D programs try to design and build IDSSs. A particular example is the ESPRIT project Graphics and Knowledge Based Dialogue for Dynamic Systems (GRADIENT) which addresses the problems of: (1) knowledge-based alarm handling with real-time state and fault identification, (2) prevention of incidents through monitoring of operator actions, and (3) increased flexibility of the graphics interface through abandoning the restrictions of pre-defined displays. The GRADIENT project regards the interaction between man and machine as a whole and integrates several specialised support functions through the common concept of a knowledge-based, graphical dialogue. The starting point is the analysis of the functionality needed for this purpose. The knowledge-based systems provides the content of the interaction, graphical expert systems provides the form, and the whole is controlled by a dialogue system. There are two categories of users for the GRADIENT system: the designer of a particular application, and the operator who uses it to control the process. In both cases the GRADIENT project considers the functions that are needed to support these user categories.