AI-driven Design

How AI can be applied in Design Tasks?  

AI can be applied to design steps in various ways. Here’s an overview: 

1. Design Generation

• Creative Design Generation (Generative AI)

2. Design Evaluation

• Performance Prediction  (Prediction Models, Physics-informed Neural Network)

• Analysis (Explainable AI)

3. Design Optimization

• Design Parameter Optimization (Reinforcement Learning, Inverse Modeling, Optimization Algorithms)

• Topology Optimization (Reinforcement Learning, Inverse Modeling, Optimization Algorithms)

Benefits of Using AI in Design Tasks

1. Design Generation - Creative Design Generation

• Generate multiple innovative design options in a fast way

• Integrated with subsequent evaluation and optimization algorithms, enabling iterative refinement and application throughout the design process 

2. Design Evaluation (Performance Prediction & Analysis)

• Evaluate faster than the simulation-based method. 

• Helps better explain why the deep learning model produces specific prediction results

3. Design Optimization (Design Parameter Optimization & Topology Optimization)

• Reduced computational cost, especially for high-dimensional problems

• Rapid adaptation to evolving design requirements 

My AI-driven Design Research

Reference: 

[1] J. T. Allison, M.-A. Cardin, C. McComb, M. Y. Ren, D. Selva, C. Tucker, P. Witherell, and Y. F. Zhao, "Artificial intelligence and engineering design," Journal of Mechanical Design, vol. 144, no. 2, p. 020301, 2022. [Online]. Available: https://doi.org/10.1115/1.4053111 
[2]  N. Yüksel, H. R. Börklü, H. K. Sezer, and O. E. Canyurt, "Review of artificial intelligence applications in engineering design perspective," Engineering Applications of Artificial Intelligence, vol. 118, p. 105697, 2023. [Online]. Available: https://doi.org/10.1016/j.engappai.2022.105697