Session: DTM-03: Computational Methods in Design

Paper Number: 147706

147706 - An Algorithm for the Optimization of Flowdown Requirements 

Common practice in the design of electro-mechanical systems is to specify the system by a set of system-level requirements. Then, as a convenient management tool, the system-level requirements are allocated (that is, “flowed down”) to the subsystem level, and thence on down to component and part levels. This process, referred to as “requirements flowdown,” allows the assignment of tasks down to design teams or even to the level of individual engineers, thus enabling concurrent design of the system. However, recognizing flowdown requirements as arbitrarily assigned constraints, we see that this process has the potential to impose significant penalties on the final system performance. It has been shown that the imposition of optimality conditions on the assignment of flowdown requirements can assure that these requirements do not penalize system performance. The optimality conditions appear to be quite simple, but their implementation presents significant problems. The objective of this research is to present a method for the optimal determination and assignment of flowdown requirements.  It is obvious that any method will require an iterative approach.  But this is difficult because we cannot write an objective function.  The function itself must be determined via an iterative design process, thus making typical search techniques such as Newton’s method or gradient search excessively expensive.  We thus developed a search technique that makes use of meta-information, that is, information about the topology of the objective function that derives from the nature of the design process rather than the results of the design process.  The algorithm that resulted from our work enables a modified gradient search that behaves more like Newton’s method in that it solves for each iteration as though it were the optimal solution.  This work moves us one large step closer to implementation in an actual case. It makes contributions to the theory of engineering design and systems engineering, which are critical to the effective design of practically all technologically complex products, and it will enable improved engineering education.  It also presents a new heuristic for optimization of design processes.

Presenting Author: Melchizedek Essandoh George Mason University

Presenting Author Biography: Melchizedek Essandoh is a graduate student at George Mason University, where he obtained his Bachelor's degree in Mechanical Engineering and is now working towards a Master's degree in Systems Engineering. His interest in engineering design research began in Fall of 2022 when he partook in an independent study working with Dr. Hazelrigg on decision-making in engineering design. He learned about the importance of implementing optimization strategies to ensure that flowdown requirements do not penalize the performance of a system. Inspired by this, he has continued to explore the optimization of flowdown requirements, looking at different approaches to improve the feasibility of the proposed technique. When he is not busy working on engineering-related problems, he enjoys playing the piano and spending time with family.

https://www.linkedin.com/in/melchizedek-e/

Authors:

Melchizedek Essandoh George Mason University
George Hazelrigg George Mason University
Jessica Nelson George Mason University