ICE: Information center expert: A consultation system for resource allocation

This paper demonstrates the viability of a rule-based consultation system as a mechanism for effective resource management through integration of knowledge about users, business problems, and resources. The specific domain within which the expert system will be tested is the Information Center (IC), which deals with end-user computing resources. In the last decade, the information center concept has been proposed repeatedly as an organizational solution to resource management problems related to end user computing. The research hypothesis is that the knowledge and methodologies of IC consultants, as well as institutional policies, can be represented in a knowledge base. The system will then draw conclusions about appropriate software or training solutions based on the interaction of user and resource profiles with the problem definition. The output of the model should be identical to the situation which it is modeling.Currently, a rule-based ES is being developed at the University of Arizona Department of Management Information Systems. This paper will present methodologies for design and implementation of the system. The development approach for the IC environment is unique as compared to expert system development as discussed in the literature. First, the use of prototyping has been stressed for expert system development with the view of using it as a throw-away system. The prototyping of an ES for the IC environment, however, has to take into account the difference between two aspects of a prototype: user interface (dialogue) and system structures (knowledge representation, inferencing techniques). There needs to be iteration of the prototype until the user if comfortable with the dialogue. The prototype will be used to extract more knowledge form the experts as well as for enhancing user dialogue. The system structures will be changed only when the user dialogue has been settled.Second, because of changes in tool availability, resulting in a need for new descriptions of the tools available, the design of an expert system for ICs has to respond dynamically to unique flexibility, portability, and maintenance issues. The ability to transport such a system to similar, but not identical, ICs should be addressed in the design as well.An IC has been described as an organization specifically designed to produce guided services to help end users help themselves. A consultation expert system might be useful in reaching that goal.