A study of new techniques for managing conflict and uncertainty in computer-aided design and engineering (CAD/E)
Abstract
Many of the computer tools which are widely used today are based on quantitative analysis and analytical models. In contrast, the basis for the systems examined in this work remain qualitative and symbolic, allowing emphasis on less defined problems in design. This work first provides an overview of existing expert CAD systems from the literature, and discusses the potential for several new techniques for managing conflict and uncertainty in a CAD/E environment. As a first example, a symbolic logic system for reliability-based design is developed. Secondly, a system for selection of rolling element bearings using the Dempster-Shafer Theory is presented. Next, a CAD system for detailed layout design of process manufacturing facilities using fuzzy optimization is described. Finally, efforts to automate preliminary design using artificial neural networks are discussed, and conclusions regarding appropriate application of each of these new techniques are made.
Degree
Ph.D.
Advisors
Rao, Purdue University.
Subject Area
Mechanical engineering|Computer science|Engineering|Artificial intelligence
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.