Research Initiation Award: System Reduction Strategies for Efficient Design Synthesis

CE设计方法要求所有工程、生产和制造团队同时互动，从而降低成本，更快地将更高质量的产品推向市场。使CE难以在工业环境中实际实现的一个难题是，由于参与组的数量过多而产生的过多的计算需求。这些群体之间的“通信路径”（可以认为是“耦合”）的数量是如此之大，以至于不可能应用正式的优化方法。研究了基于灵敏度的策略，在不牺牲解决方案准确性的情况下，减少参与设计小组之间所需的通信路径的数量。这些策略通过人工智能技术得到增强，以实现人机界面功能。重新设计设计过程本身是为了促进并行工程实现。解决的一些研究问题包括：子系统之间耦合强度的量化，暂停或消除候选耦合的识别，以及最优收敛方案。美国的汽车、航空航天、计算机和电子工业等被迫对竞争压力作出反应，由于产品开发时间长和成本过高，导致市场份额的严重损失。这些行业已经认识到采用并行工程（CE）设计方法的必要性，以便以更低的成本和时间创建更高质量的产品。本研究的目的是开发人工智能增强策略，以提高大规模复杂工程系统的设计过程效率，例如在并行工程（CE）应用中遇到的问题。复杂工程系统的特点是参与设计群体之间存在固有的耦合，这使得传统的设计和优化方法变得不切实际。

The CE design approach requires the interaction of all engineering, production, and manufacturing groups simultaneously so as to reduce cost and bring higher quality products to market faster. One dilemma that makes CE difficult to practically implement in an industrial setting is the excessive computational requirements due to the inordinate number of participating groups. The number of 'communication paths' amongst these groups (which can be considered as 'couplings') is so large that formal optimization methods are impossible to apply. Sensitivity-based strategies that reduce the number of required communication paths amongst participating design groups, without sacrificing solution accuracy, are researched. These strategies are augmented by AI techniques to allow for human interface capabilities. A redesign of the design process itself is so as to facilitate concurrent engineering implementation. Some of the research issues addressed include: quantification of coupling strengths between subsystems, identification of candidate couplings for suspension or elimination, and optimal convergence schemes. United States automotive, aerospace, computer and electronics industries, among others, have been forced to respond to competitive pressures that have resulted in crippling losses of market share due to extensive product development times and inordinate costs. These industries have recognized the necessity of adopting the Concurrent Engineering (CE) design approach in order to create higher quality products that can be brought to market for significantly lower cost and time. The objective of this research is to develop AI-augmented strategies for increasing the efficiency of the design process for large scale, complex engineering systems, such as are encountered in Concurrent Engineering (CE) applications. Complex engineering systems are characterized by inherent couplings amongst the participating design groups, which makes traditional design and optimization approaches impractical.