This study proposes a constraint-setting support method for the early stages of design, where uncertainty is high. The goal is to improve the efficiency and accuracy of design space exploration by appropriately narrowing down the dataset consisting of design variables and performance values that serve as candidates for potential design solutions. Specifically, we comprehensively extract constraint patterns applied to performance values and derive the corresponding design spaces that satisfy each constraint. These design spaces are then evaluated using metrics that represent ease of design and performance preference. Finally, the evaluation results are visualized in a manner that is easy for designers to understand, thereby supporting them in setting appropriate constraints. The proposed method is applied to the high-dimensional design problem of an electric vehicle (EV) motor to demonstrate its effectiveness. As a result, the method enabled the setting of constraints that reflect the designer’s intent, while achieving both higher ease of design and higher performance preference.
Nawa et al. (Wed,) studied this question.