UBC Theses and Dissertations
A framework for qualitative and semi-quantitative analysis in engineering design and evaluation Gedig, Michael H.
A framework is presented which is used to support engineering decision-making in the presence of incomplete and imprecise information. A number of sound reasoning techniques which utilize this framework are described. These techniques include qualitative reasoning, a branch of study in the field of artificial intelligence, and semi-quantitative reasoning, which draws inferences from partially-specified numeric values. The constraint-satisfaction paradigm plays an important role in the implementation of these reasoning strategies. The notion of the constraint is well-suited to engineering planning and design, where projects are generally formulated in terms of a set of functional specifications. A computer program was developed based on the framework presented here. The program implements several complementary qualitative and semi-quantitative reasoning strategies. The specific techniques used in the program include constraint-satisfaction algorithms for both qualitative and numeric domains, interval arithmetic, and constraint inference methods. It was intended that the computer program developed as a part of this research serve as a practical tool which could provide logical justification for engineering decisions. For this reason, the program was developed on a platform which is both accessible and familiar. As part of the program, a convenient method of systematically specifying constraints was created, which mirrors the way in which engineering problems are normally specified. Examples are given which illustrate the application of the techniques presented here. These examples show that the techniques discussed in this thesis have significant potential for justifying engineering decisions made under conditions of uncertainty.
Item Citations and Data