EAGER: Advancing Understanding of Function Modeling, Reasoning, and Thinking in Design by Mapping Behaviors to Cognitive Explanations

Clemson University, Clemson SC

Investigators

Abstract

In engineering design, a popular representation taught in mechanical design curricula is "functional modeling." Function has been defined as what a device or system is intended to do or the intentional transformation from inputs to outputs. The function is not what the product is (form), how the product does what it does (behavior), or why the product exists (purpose). However, it does relate to all of these, and thus allows engineers to transition between the problem space and the solution space. Many design tools have been developed based on functional modeling but are not yet ubiquitously found throughout industry. Rather than suggesting new function vocabularies, tools, or methods, this EArly-concept Grant for Exploratory Research (EAGER) project seeks to understand how engineers think about function by studying their behaviors in creating models. To do this, the modeling behaviors of design engineers will be recorded, and models of cognition will be used to explain these behaviors. With this understanding, new computational design tools can be developed in the future. In addition, engineering design education can be enhanced by providing students with better strategies for creating the function models. It is important to understand how engineers think, reason, and operate on function modeling elements and how these models bridge problem definition and solution exploration. Through experimental protocol studies, the specific modeling patterns (forward chaining, backward chaining, nucleation) and pause patterns (pause before or after elemental instantiation) will be analyzed. The modeling patterns will provide evidence of goal-oriented thinking (backward chaining) or causal reasoning (forward chaining). The pause patterns and frequencies will be used to understand relative importance of function (action of transformation) and flow (things being transformed). Two basic research questions will be explored: Are there modeling behavior patterns to function modeling that are invariant of chosen representation? And can these modeling behavior patterns be used to understand, predict, and explain cognitive models of design activities?