Integrated Parameter and Tolerance Design with Computer Experiments
DescriptionParameter and tolerance design are two important activities in quality engineering. They have been formulated as optimization problems in which values for the means (parameter design) and tolerances of input characteristics are adjusted to optimize quality and total costs. Numerous methods for parameter and tolerance design have been developed for cases where the response function that links the output characteristics and the input characteristics is known or can be estimated through physical experiments. This project aims to develop integrated parameter and tolerance design (IPTD) methodologies for problems in which the response function is a time-consuming computer code. In these cases, the function has to be estimated through computer experiments due to the time-consuming nature of the simulations. The Gaussian process model will be used to estimate and quantify uncertainty about the functional relationship. Methods for optimizing the total cost that takes into account this uncertainty will be developed.
|Effective start/end date||1/10/13 → 10/03/16|