The use of surrogate model-based optimization in engineering design has been on the rise. This method involves creating a surrogate model using data from simulations or real-world experiments, and then using numerical optimization techniques to find the optimal solution. Recent advancements in deep learning-based inverse design methods have made it possible to generate real-time optimal solutions for engineering design problems without the need for iterative optimization processes. However, there has not been a comprehensive study comparing the advantages and disadvantages of this approach to traditional design optimization methods. This paper aims to compare the performance of traditional design optimization methods with deep learning-based inverse design methods using benchmark problems in various scenarios. The results of this study will provide guidelines for the future use of deep learning-based inverse design, improving its practical applicability to real engineering design problems.
Comparing Performance of Design Optimization and Deep Learning-based Inverse Design: A Study (arXiv:2308.13000v1 [math.OC])
