Details
| Original language | English |
|---|---|
| Pages (from-to) | 609-620 |
| Number of pages | 12 |
| Journal | Journal of Zhejiang University: Science A |
| Volume | 22 |
| Issue number | 8 |
| Publication status | Published - Aug 2021 |
Abstract
We propose the deep Lagrange method (DLM), which is a new optimization method, in this study. It is based on a deep neural network to solve optimization problems. The method takes the advantage of deep learning artificial neural networks to find the optimal values of the optimization function instead of solving optimization problems by calculating sensitivity analysis. The DLM method is non-linear and could potentially deal with nonlinear optimization problems. Several test cases on sizing optimization and shape optimization are performed, and their results are then compared with analytical and numerical solutions.
Keywords
- Artificial neural networks, Deep learning, Sensitivity analysis, Structural optimization, TP183, TU31
ASJC Scopus subject areas
- Engineering(all)
- General Engineering
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Journal of Zhejiang University: Science A, Vol. 22, No. 8, 08.2021, p. 609-620.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - A deep neural network-based algorithm for solving structural optimization
AU - Kien, Dung Nguyen
AU - Zhuang, Xiaoying
N1 - Funding Information: The authors would like to express the appreciation to Prof. Dr. Krister SVANBERG (KTH Royal Institute of Technology, Sweden) for his MMA codes, and Prof. Dr.-Ing. Timon RABCZUK (Bauhaus-Universit?t Weimar, Germany) for his critical comments on the manuscript. Funding Information: Project supported by the ERC StG (No. 802205)
PY - 2021/8
Y1 - 2021/8
N2 - We propose the deep Lagrange method (DLM), which is a new optimization method, in this study. It is based on a deep neural network to solve optimization problems. The method takes the advantage of deep learning artificial neural networks to find the optimal values of the optimization function instead of solving optimization problems by calculating sensitivity analysis. The DLM method is non-linear and could potentially deal with nonlinear optimization problems. Several test cases on sizing optimization and shape optimization are performed, and their results are then compared with analytical and numerical solutions.
AB - We propose the deep Lagrange method (DLM), which is a new optimization method, in this study. It is based on a deep neural network to solve optimization problems. The method takes the advantage of deep learning artificial neural networks to find the optimal values of the optimization function instead of solving optimization problems by calculating sensitivity analysis. The DLM method is non-linear and could potentially deal with nonlinear optimization problems. Several test cases on sizing optimization and shape optimization are performed, and their results are then compared with analytical and numerical solutions.
KW - Artificial neural networks
KW - Deep learning
KW - Sensitivity analysis
KW - Structural optimization
KW - TP183
KW - TU31
UR - http://www.scopus.com/inward/record.url?scp=85113435751&partnerID=8YFLogxK
U2 - 10.1631/jzus.A2000380
DO - 10.1631/jzus.A2000380
M3 - Article
AN - SCOPUS:85113435751
VL - 22
SP - 609
EP - 620
JO - Journal of Zhejiang University: Science A
JF - Journal of Zhejiang University: Science A
SN - 1673-565X
IS - 8
ER -