Qualitative Reasoning in an Expert System Framework.
Stephen Edward Cross · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · 1983
Abstract : Explanation capabilities of an expert system in the domain of air traffic control is examined. Qualitative reasoning is used to justify heuristically generated plans. The justification is based on mathematical knowledge of aircraft performance which is computationally too complex for use in the normal planning process. Equations are represented in a semantic network which provides a structure in which qualitative values are propagated. The approach is unique in three aspects. First, an abstraction level is used. The equations can be interpreted in terms of Newtonian mechanics as applied to the one dimensional motion. Second, the reasoning process is bidirectional. Third, the computer generates the semantic structure based on a symbolic series expansion of the equations. (Author)