Rule-Based Versus Structure-Based Models for Explaining and Generating Expert Behavior

Flexible representations are required in order to understand and generate expert behavior.While production rules with quantifiers can encode experiential knowledge, they often haveassumptions implicit in them, making them brittle in problem scenarios where theseassumptions do not hold. Qualitative models achieve flexibility by representing the domainentities and their interrelationships explicitly. However, in problem domains whereassumptions underlying such models change periodically, it is necessary to be able to synthesizeand maintain qualitative models in response to the changing assumptions. In this paper, weargue for a representation that contains partial model components that are synthesized intoqualitative models containing entities and relationships relevant to the domain. The modelcomponents can be replaced and rearranged in response to changes in the task environment.We have found this quot;model constructorquot; to be useful in synthesizing models that explain andgenerate expert behavior, and have explored its ability to support decision-making in theproblem domain of business resource planning, where reasoning is based on models that evolvein response to changing external conditions or internal policies