A Computational Approach to Engineering Idea Generation

生成工程创意的计算方法

• 批准号: 8618776
• 负责人: Warren Seering
• 金额: --
• 依托单位: Massachusetts Institute of Technology
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1987
• 资助国家: 美国
• 起止时间: 1987-01-01 至 1988-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8618776&HistoricalAwards=false
• 关键词: Computational; Approach; Engineering; Idea; Generation

It is proposed to test the hypothesis that physical descriptions of new devices can be synthesised from knowledge about the physical and functional characteristics of analogous devices. The approach builds upon concepts developed in the area of machine learning. Research on learning to categorize objects by knowing a physical description and inferring a functional definition has provided the insight to use examples to learn a physical description from a functional definition. This research will concentrate on generating solutions to defined problems in the mechanical design domain. Progress in this area could make a significant contribution to understanding conceptual design.

本文提出了一种假设，即新器件的物理描述可以从类似器件的物理和功能特征的知识中合成。该方法建立在机器学习领域开发的概念之上。通过了解物理描述和推断功能定义来学习对对象进行分类的研究提供了使用示例从功能定义中学习物理描述的洞察力。这项研究将集中在生成解决方案，以确定在机械设计领域的问题。这一领域的进展将对理解概念设计做出重大贡献。