Deep Generative Models for Geometric Design Under Uncertainty
This addresses uncertainty quantification in engineering design, though it appears incremental as it extends existing deep generative models with conditional distributions.
The paper tackles the problem of geometric design optimization under manufacturing uncertainty by proposing GAN-DUF, a framework that learns both nominal designs and conditional distributions of fabricated designs, demonstrating improved performance in two real-world engineering examples.
Deep generative models have demonstrated effectiveness in learning compact and expressive design representations that significantly improve geometric design optimization. However, these models do not consider the uncertainty introduced by manufacturing or fabrication. Past work that quantifies such uncertainty often makes simplified assumptions on geometric variations, while the "real-world" uncertainty and its impact on design performance are difficult to quantify due to the high dimensionality. To address this issue, we propose a Generative Adversarial Network-based Design under Uncertainty Framework (GAN-DUF), which contains a deep generative model that simultaneously learns a compact representation of nominal (ideal) designs and the conditional distribution of fabricated designs given any nominal design. We demonstrated the framework on two real-world engineering design examples and showed its capability of finding the solution that possesses better performances after fabrication.