Evaluating Qualitative Reasoning as a Support Tool for the Transfer of Conceptual Knowledge

Samenvatting

Qualitative reasoning (QR) is the area of Artificial Intelligence (AI) that captures and simulates conceptual knowledge about system behaviour. This thesis explores how QR can support the transfer of conceptual knowledge and how tools that provide that support can be evaluated. The approach in this thesis is to create an evaluation framework. This framework provides structure for exploring and evaluating how a QR software program can support knowledge transfer. The framework contains five types of support QR can provide to knowledge transfer. Dimensions are provided to describe usages of these types and guidance for evaluation those usages are set up. To investigate the usefulness of this framework it applied to Garp3. Garp3 is a software tool with a graphical interface that can be used to build and simulate QR models. Two studies are conducted to evaluate two usages. The first study addresses the acquisition of conceptual knowledge by working with QR and showed that novice users can work and learn with Garp3 with minimal software instructions. The second study addresses the articulation of conceptual knowledge with a recently developed structured approach. This study showed that novice users were able to build models with Garp3 with minimal software instructions. However, instruction and support to the structured approach is vital to model building success. The thesis discusses QR, ways to use and evaluate tools based on QR, and the specifics to the two evaluations studies that have been carried out.