EAGER: Theoretical Foundation for the Use of Approximations and Heuristics in Systems Engineering and Design

Georgia Tech Research Corporation, Atlanta GA

Investigators

Abstract

In practice, engineers often rely on heuristics and approximations to arrive at good system designs efficiently and effectively. Heuristics are rules of thumb that embody engineering experience. When used appropriately, they save engineering time and effort without sacrificing the quality of the system being engineered. However, the use of heuristics and approximations is poorly understood. This EArly-concept Grant for Exploratory Research (EAGER) project will enable a fundamental understanding and development of design heuristics and associated processes, guiding researchers and practitioners in advancements and use of such heuristics, and ultimately, providing a formalized foundation to design science. The research will combine mathematical foundations, such as probability theory and decision theory, with empirical observations of students and experienced engineers to develop explanatory models for the use of heuristics and approximations. The research will make strides in formalizing design heuristics from various disparate approaches, and recommend how and when heuristics are best used in design practice. In addition, it will lead to new insights regarding how to teach design better by clearly identifying the heuristic nature of design knowledge, and by coaching students to continually update their heuristics through the contextualization of the underlying foundations and through life-long learning. Systems engineering and design currently lack a strong theoretical foundation, and, in particular, the community has struggled to bridge the gap between current practice and the normative, axiomatic, mathematical theories that underlie systems engineering and design. To fill this gap in knowledge, the objective of this research is to enhance the scientific understanding of the use of approximations and heuristics in systems engineering and design. A normative theory will be derived for the use of approximations and heuristics by explicitly accounting for the value of the resources required to execute designer actions, starting from the von Neumann-Morgenstern axioms of decision theory. Heuristics found will be cataloged and analyzed by characterizing their value or 'validity.' The relationship between design heuristics and cognitive heuristics as studied in psychology will be considered, bridging the gap between descriptive and prescriptive theories, identifying gaps between the two, and deriving hypotheses regarding the improvement of design practice in situations in which the gap between normative theory and current practice appears to be large. This research will assist in enabling a fundamental understanding and development of design heuristics and associated processes, and will guide researchers and practitioners in advancements and use of such heuristics, and ultimately, provide formalized foundations to design science.