Representation and control of knowledge bases for support of multiple tasks

The goal was to develop a use-independent knowledge structure at the domain and strategy level and to enable a knowledge-based system to exhibit diverse dimensions of expertise, such as problem-solving, explanation, and learning, using the knowledge base. The authors show the functional capabilities of such advanced generic expert systems with respect to representation and control of problem solving strategy knowledge. In this approach, domain and meta level knowledge is represented in a declarative, explicit, and modular way. This improved representation at the strategy and domain level enables the performance system, MINERVA, and the learning program, ODYSSEUS, to use the same knowledge base. Explicit representation of schedular knowledge enables MINERVA to solve a problem opportunistically and to generate multi-level explanations of its own problem-solving