Engineering Analysis with Uncertainties and Complexities, Using Reasoning Approaches

Abstract

Conventional computation methods generally limit practicing engineers from using complex formulations or considering uncertainties in general. A method is needed that can be implemented regardless of the uncertainty or linearity of the design parameters and their constraints. Methods such as qualitative reasoning provide an effective and sound technique for solving complex and uncertain scenarios. Uncertainties in engineering designs can be formulated as variables in the application domain and processed by numerical constraint reasoning. This paper describes the theories and algorithms behind a software platform built upon numerical constraint reasoning for engineering applications. The capability of representing design parameters and outcomes in a 2D solution space provides a practical way for engineers to leverage their existing knowledge and experience. The software expresses the results of the analysis in variable ranges and diagrams showing a 2D design space. Qualitative reasoning can assist in the difficult process of making appropriate engineering assumptions and judgments when carrying out complicated analysis procedures. In addition, interval constraint analysis can be used to derive controlling parameters and design space, therefore giving engineers a good overall understanding of a problem when practical experience is not available.