Abstract

Engineering Design represents a highly creative process that is decision rich. Due to the immature nature of the design information during the earliest phases of the design process, these decisions are frequently taken with incomplete or uncertain information leading to risk exposure. Engineering designers use their own subjective experience to direct them through this process. This seminar explores support methodologies that are able to inform his decision making process and thus propose a conceptual Bayesian based decision framework. Based on the literature, two variants of this framework are presented: one which will lead the designer ‘down the path of least resistance’ and one that will encourage creativity.

Biography: Peter Matthews is a Lecturer in Design Informatics in the School of Engineering at the University of Durham. His core research interests lie in applying Machine Learning techniques to design problems. In particular, he is interested in extracting the design knowledge learnt by the machine. This extracted knowledge can then be used to guide the documentation process for a design domain. This is particularly useful in domains that are not fully understood or that do not have analytical models, eg. aesthetic evaluation.