DATA-DRIVEN CREATIVITY: COMPUTATIONAL PROBLEM-EXPLORING IN ENGINEERING DESIGN



Obieke, Chijioke C, Milisavljevic-Syed, Jelena and Han, Ji ORCID: 0000-0003-3240-4942
(2021) DATA-DRIVEN CREATIVITY: COMPUTATIONAL PROBLEM-EXPLORING IN ENGINEERING DESIGN. In: 23rd International Conference on Engineering Design, 2021-8-16 - 2021-8-20, Gothenburg.

[img] Text
Data Driven Creativity - ELE.docx - Author Accepted Manuscript

Download (3MB)

Abstract

<jats:title>Abstract</jats:title><jats:p>Creativity is required in engineering design. It is required in the aspects of problem-solving - conceptualizing a new solution to a problem, and problem-exploring - conceptualizing a new problem. Studies show that, in both aspects, creativity is a difficult task in practice. The aim of this study is to support the engineering design community by easing the difficulty in the problem-exploring practice. To achieve this, a computational problem-exploring (CPE) model is developed to mimic how design engineers identify a valid design problem. Consequently, a CPE tool - Pro-Explora V1 is developed based on the CPE model. The CPE model consists of a synergy of emergent computational technologies including data retrieval and machine learning. A Markovian model is employed in the CPE model to enable a data-driven random process for exploring design problems. In pilot test, Pro-Explora V1 generated some engineering design-related problems which are meaningful, unique, and could not be distinguished from naturally generated ones. It provides support to design engineers in problem-exploring at the early stage in engineering design. This study contributes to the global effort towards data-driven processes in the fourth industrial revolution.</jats:p>

Item Type: Conference or Workshop Item (Unspecified)
Uncontrolled Keywords: Bioengineering
Divisions: Faculty of Science and Engineering > School of Engineering
Depositing User: Symplectic Admin
Date Deposited: 10 Jun 2021 07:41
Last Modified: 15 Mar 2024 12:18
DOI: 10.1017/pds.2021.83
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3125742