Trade-Off Based Robust Modeling and Design

这笔赠款提供资金，用于开发基于权衡的稳健建模和设计方法，以确定统计上最优的产品规格。 这项研究将利用从效用理论到一个强大的设计范式的概念，定量地将定性知识和不同属性的偏好，而不失一般性和准确性。 将开发基于统计探索的实验设计技术，以探索和生成系统行为信息，并从属性数据的定量表示中确定最佳产品规格。 通过在设计过程中结合更高层次的模型和设计知识的显式表示，将开发一个交互式设计系统，使设计师在设计过程中做出明智的决定。 行业驱动的设计问题将被用作案例研究，以测试和评估所提出的方法。 如果成功，这项研究将导致一个统一的方法，提供了一个严格的处理设计过程中，从整体设计的角度在数据的不确定性条件下。 本研究的主要目标是确定一个权衡为基础的决策模型制定的直接和同时治疗的多个目标和约束的设计过程中，其集成与统计探索为基础的强大的最优设计生成策略。 这项研究可能会导致更好地理解工程建模过程，并推进知识的状态，所产生的固有的复杂性表示物理设计问题，使用理想化的基于计算机的抽象可以得到解决。 本研究的结果也将有助于确定一个一致的机构的协同，综合工程设计方法的基础上设计的实验，实用理论，基于计算机的仿真模型，有限元方法和设计优化原则。

This grant provides funding for the development of a trade-off based robust modeling and design methodology for identification of statistically optimal product specifications. This research will exploit concepts from utility theory into a robust design paradigm to quantitatively incorporate qualitative knowledge and preferences of different attributes without loss of generality and accuracy. Statistical exploration based design of experiment techniques will be developed to explore and generate system behavioral information, and to identify optimal product specifications from this quantitative representation of attribute data. By incorporating explicit representation of higher level modeling and design knowledge in the design process, an interactive design system will be developed to enable designers make intelligent decisions during the design process. Industry-driven design problems will be used as case studies to test and evaluate the proposed methodology. If successful, this research will result in a unified methodology that offers a rigorous treatment of design process from an overall design perspective under conditions of uncertainty in data. The primary goal of this research involves the determination of a trade-off based decision model formulation for direct and simultaneous treatment of multiple objectives and constraints in the design process, and its integration with a statistical exploration based robust optimal design generation strategy. This research could lead to a better understanding of the engineering modeling process, and advance the state of knowledge by which the inherent complexities arising from representing physical design problems using idealized computer-based abstractions can be addressed. Results of this research will also contribute towards the identification of a consistent body of synergistic, integrated engineering design methods based on design of experiments, utility theory, computer-based simulation models, finite element methods, and design optimization principles.