Collaborative Research: Creating a Computational Theory for Conceptual Design

这个项目结合了新的设计表示和计算方法，使计算概念设计成为可能。设计表示和计算方法利用不断增长的计算能力和信息检索推理算法。描述方法、存储知识和算法推理为概念设计的计算理论提供了潜力，可以在概念生成的困难阶段极大地造福设计者。潜在的主题是基于功能的综合、约束管理和状态空间搜索的组合和形式化。这项工作的主要目标是创建一个概念设计的计算理论，可以计算设计方案。计算将得到一个概念变体的综合空间，并在其中搜索可行的候选空间。概念生成活动是工程设计的基石之一。直到最近，设计师在概念设计过程中唯一可用的资源是个人经验和天生的能力。虽然设计师的资源在过去三十年中有了显著的进步，但计算设计工具和概念设计方法之间仍然缺乏连续性。许多概念设计的形式化方法尚未实现为计算算法。这项研究将开发表征和计算方法，允许从所需产品的功能描述中创建设计概念。通过开发设计知识库和使用存储知识的概念生成器，本研究扩展了当前对功能与形式关系的理解，并将其编纂。此外，新的组件表示使计算机能够确定组件兼容性，从而合成概念变体。此外，这项工作对于与人类设计行为进行比较很有用，因为它是建立在常用的设计实践之上的。这项工作是基础性的，也是实践性的。因此，更广泛的影响超越了出版，延伸到在教育、工程实践和研究中使用所得到的知识、方法和工具。例如，学生将能够使用概念生成器来创建解决方案，因为他们的经验有限，否则他们可能无法开发。这些解决方案将使学生接触到不同的学科、工程和科学领域。这项研究的结果以类似的方式影响了工业和研究。在这两种情况下，这项研究都可以用来产生更广泛的解决方案，比没有它要快得多。

This project combines new design representations and computational approaches to enable computational conceptual design. The design representations and computational methods leverage the constant increases in computational power and information retrieval reasoning algorithms. Descriptive methods, stored knowledge, and algorithmic reasoning provide the potential for a computational theory of conceptual design that could greatly benefit the designer during difficult stages of concept generation. The underlying theme is the combination and formalization of function-based synthesis, constraint management and state space search. The primary objective of the work is to create a computational theory of conceptual design that can compute design alternatives. The computation will result in a comprehensive space of concept variants and search it for feasible candidates.The activity of concept generation is one of the cornerstones of engineering design. Until recently, the only resources available to a designer during conceptual design were personal experiences and innate abilities. While the designer's resources have advanced significantly in the last three decades, there is still a lack of continuity between computational design tools and conceptual design methods. Many formal methods of conceptual design have yet to be realized as computational algorithms. This research will develop representations and computational methods that allow design concepts to be created from the functional description of a needed product. Through the development of a design knowledge repository and a concept generator that uses the stored knowledge, this research extends the current understanding of the relationship between function and form and codifies it. Also, novel component representations enable computers to determine component compatibility and thus synthesize a concept variant. Additionally, this work is useful for comparison to human design behaviors, as it is built on commonly used design practices.This work is both basic and practical in nature. Thus, the broader impacts extend beyond publication to the usage of the resultant knowledge, methods and tools in education, engineering practice and research. For example, students will be able to use the concept generator to create solutions that, because of their limited experience, they might otherwise not have developed. These solutions will expose students to different disciplines and engineering and scientific domains. The results of this research impact industry and research in a similar manner. In both cases, the research can be used to generate a broader array of solutions much more quickly than would have been possible without it.