EAGER: A Fine-Grained Data-Driven Approach to Studying Sequential Decision-Making in Engineering Systems Design

EAGER：一种研究工程系统设计中顺序决策的细粒度数据驱动方法

• 批准号: 1842588
• 负责人: Zhenghui Sha
• 金额: $ 22.5万
• 依托单位: University of Arkansas
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-09-01 至 2021-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1842588&HistoricalAwards=false
• 关键词: EAGER; Fine; Grained; Data; Driven

This EArly-concept Grant for Exploratory Research (EAGER) grant supports fundamental research into sequential decision making in engineering systems design. Each decision an engineer makes during a design process impacts the direction and outcomes of the project. An improved understanding this process can lead to better guidelines and support tools for engineering designers and, in turn, improved engineered systems and overall industrial competitiveness. However, an important challenge in studying these processes is the difficulty of obtaining fine-grained empirical data of engineering designers in action. This project will create and demonstrate a research platform for the large-scale data acquisition and analysis of decision processes in engineering systems design. This new research approach provides a high-resolution lens for probing into design thinking and will enable researchers to identify design thinking patterns and strategies that are not evident through other observational techniques. This can lead to valuable insights that have a major impact on engineering design education, practitioner strategies, and engineering tools. Specific outcomes of this project include the creation of the open-source fine-grained data-driven research platform, dissemination of the platform to other researchers, and demonstration of its use to investigate engineering design thinking through empirical studies of systems thinking and sequential decision making in the design of solar energy systems. The primary objective of this high-risk high-reward project is to create and demonstrate a research approach centered on the acquisition and analysis of fine-grained design activity data for design research. The approach is based on an open-source research experiment platform extended from an existing computer-aided design (CAD) software, Energy3D, for renewable energy systems design. This project will 1) extend Energy3D to incorporate functionality required for a research platform, 2) demonstrate use of the new research platform to support the acquisition of fine-grained data from real-world design exercises, and 3) disseminate the platform within the engineering design research community through publications and tutorials. The research study will highlight how fine-grained data enables new research directions on sequential decision-making and system thinking, two fundamental elements of engineering design thinking. Specifically, the approach combines Markov decision process and deep neural networks with data from human-subject experiments to establish decision process models. A principal risk of this project is that there may be limits to the conclusions researchers can draw based primarily on the observed actions of designers. However, the potential reward is deep insight into designers' sequential decision-making and its interaction with systems thinking. This is expect to lead to recommendations for improved engineering design strategy and transformative next-generation design tools.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

这项早期概念探索性研究资助 (EAGER) 资助支持工程系统设计中顺序决策的基础研究。工程师在设计过程中做出的每个决定都会影响项目的方向和结果。更好地理解这一过程可以为工程设计师提供更好的指导和支持工具，进而提高工程系统和整体工业竞争力。然而，研究这些过程的一个重要挑战是难以获得工程设计师实际操作的细粒度经验数据。该项目将创建并演示一个用于工程系统设计中决策过程的大规模数据采集和分析的研究平台。这种新的研究方法为探索设计思维提供了高分辨率的镜头，并使研究人员能够识别通过其他观察技术不明显的设计思维模式和策略。这可以带来宝贵的见解，对工程设计教育、从业者策略和工程工具产生重大影响。该项目的具体成果包括创建开源细粒度数据驱动研究平台，将该平台传播给其他研究人员，并通过对太阳能系统设计中的系统思维和顺序决策的实证研究，展示其用于研究工程设计思维的用途。这个高风险高回报项目的主要目标是创建和演示一种以获取和分析用于设计研究的细粒度设计活动数据为中心的研究方法。 该方法基于一个开源研究实验平台，该平台从现有的计算机辅助设计 (CAD) 软件 Energy3D 扩展而来，用于可再生能源系统设计。该项目将 1) 扩展 Energy3D 以纳入研究平台所需的功能，2) 演示如何使用新的研究平台来支持从实际设计练习中获取细粒度数据，以及 3) 通过出版物和教程在工程设计研究社区内传播该平台。该研究将强调细粒度数据如何为顺序决策和系统思维（工程设计思维的两个基本要素）的新研究方向提供支持。具体来说，该方法将马尔可夫决策过程和深度神经网络与人体实验数据相结合，建立决策过程模型。该项目的主要风险是，研究人员主要根据观察到的设计师行为得出的结论可能存在限制。然而，潜在的回报是深入洞察设计师的顺序决策及其与系统思维的相互作用。预计这将带来改进工程设计策略和变革性下一代设计工具的建议。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Developing Instructional Design Agents to Support Novice and K-12 Design Education

开发教学设计代理以支持新手和 K-12 设计教育

• 发表时间: 2019
• 期刊: ASEE annual conference & exposition
• 作者: Schimpf, Corey;Huang, Xudong;Xie, Charles;Sha, Zhenghui;Massicotte, Joyce
• 通讯作者: Massicotte, Joyce

Automatic Clustering of Sequential Design Behaviors

顺序设计行为的自动聚类

• DOI: 10.1115/detc2018-86300
• 发表时间: 2018
• 期刊: ASME 2018 International Design Engineering Technical Conferences & Computers and Information in Engineering Conference
• 作者: Rahman, Molla Hafizur;Gashler, Michael;Xie, Charles;Sha, Zhenghui
• 通讯作者: Sha, Zhenghui

Integrating Sequence Learning and Game Theory to Predict Design Decisions Under Competition

• DOI: 10.1115/1.4048222
• 发表时间: 2020-10
• 期刊: Journal of Mechanical Design
• 影响因子: 3.3
• 作者: A. E. Bayrak;Zhenghui Sha

MODELLING AND PROFILING STUDENT DESIGNERS’ COGNITIVE COMPETENCIES IN COMPUTER-AIDED DESIGN

对学生设计师的计算机辅助设计认知能力进行建模和分析

• DOI: 10.1017/pds.2021.477
• 发表时间: 2021
• 期刊: Proceedings of the Design Society
• 作者: Clay, John;Li, Xingang;Rahman, Molla Hafizur;Zabelina, Darya;Xie, Charles;Sha, Zhenghui
• 通讯作者: Sha, Zhenghui

Predicting Sequential Design Decisions Using the Function-Behavior-Structure Design Process Model and Recurrent Neural Networks

• DOI: 10.1115/1.4049971
• 发表时间: 2021-08
• 期刊: Journal of Mechanical Design
• 影响因子: 3.3
• 作者: M. H. Rahman;Charles Xie;Zhenghui Sha