Collaborative Research: Validating Predictive Models in Engineering Design

协作研究：验证工程设计中的预测模型

• 批准号: 0522366
• 负责人: Kwokleung Tsui
• 金额: $ 20.74万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2005
• 资助国家: 美国
• 起止时间: 2005-09-01 至 2009-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0522366&HistoricalAwards=false
• 关键词: Collaborative; Research; Validating; Predictive; Models

The objective of this collaborative research is to develop a model validation approach that addresses the need for validating predictive models in the context of design decision making. Modeling and simulation based design has been increasingly used for designing engineering systems. However, it remains a challenge on assessing the risks and uncertainties associated with the use of predictive models in engineering design. Existing model validation approaches are limited in their scopes and lack of stochastic measurements with regard to the confidence of making specific design decisions. In the proposed research, a rigorous model validation approach will be developed to address the need of validating predictive models for engineering design. Besides, model validation metrics, statistical modeling, Bayesian inferences, data mining techniques, and decision making strategies will be developed to incorporate model validation as an integral part of design decision making.If successful, this research will offer a generic model validation approach that can be applied to many domestic and military applications for making reliable decisions with predictive models as replacements of expensive physical part deployments. The strong collaborations between the research teams, industrial partners, and government agencies will ensure that the developed technologies and results will be transferred and successfully implemented. The research will contribute to education in the areas of model-based simulation, modeling and optimization of engineering systems under uncertainty, statistical analysis, engineering design, and information technology.

这项合作研究的目的是开发一种模型验证方法，解决设计决策的背景下验证预测模型的需要。 基于建模和仿真的设计已越来越多地用于工程系统的设计。 然而，它仍然是一个挑战，评估与使用预测模型在工程设计中的风险和不确定性。 现有的模型验证方法是有限的，在他们的范围和缺乏随机测量方面的信心，使具体的设计决策。 在拟议的研究中，将开发一个严格的模型验证方法，以满足验证预测模型的工程设计的需要。 此外，我们亦会发展模型验证指标、统计模型、贝氏推论、数据挖掘技术及决策策略，使模型验证成为设计决策的一部分。如成功，这项研究将提供一种通用的模型验证方法，可以应用于许多国内和军事应用，以便用预测模型代替昂贵的物理部件来做出可靠的决策部署。 研究团队、工业合作伙伴和政府机构之间的密切合作将确保开发的技术和成果得到转让和成功实施。 该研究将有助于教育领域的基于模型的仿真，建模和优化的工程系统下的不确定性，统计分析，工程设计和信息技术。