sunnyshijuncheng
commented
2 years ago Dear Hao Xu,
Thanks much for your response. I have add you QQ, please check.
Best regards,
Shijun Cheng Postdoctoral Fellow Seismic Wave Analysis Group (SWAG) Physical Sciences & Engineering Division Phone: +966 (0)56 753 0947 Email: @.*** King Abdullah University of Science and Technology Bldg#1, Level#3, Office#3203-CU08 Thuwal 23955-6900, Kingdom of Saudi Arabia
woshixuhao @.***> 于2022年10月9日周日 11:57写道：
P.S. I am very happy to finally contact with you! I have tried to send several e-mails to your instituional mailbox, however, they are all rejected and sent back. I hope you can receieve this e-mail. I find that this reply mail cannot attach attachments, so you can add my QQ: 390260267 for further contact.
Hello, Thanks for your interest of our works! The code is attached in this e-mail. Here, I provide the case of parametric-convection-diffusion to help you better try this code. I am sorry that the code have not been sorted out well, so it may takes some time to get familiar with it.
The code is run as mentioned below:
run main.py to train a neural network, and the settings of network is in the neural_network.py
run GA-produce-meta-data.py to obtain the meta-data and then run 结合遗传PDE.py to conduct DLGA(structure). Here, you may manually adjust the domain of meta-data (line 16 and 17) to obtain the meta-data in different local windows and then run the 结合遗传PDE.py get the structure in this windows. After obtaining the structure in all windows, the optimal structure can be obtained as illustrated in the article.
run adaptive.py to get the coefficient of each term in each x or t. (This is a easier and faster way to obtain the coefficient, and the article use the PINN to optimize, but it is slow, if you want the PINN code, please tell me.)
run new-变系数.py to conduct DLGA(coefficient).
The middle result have been provided in the code, and you can debug anywhere. If you have any question ,please contact me at any time!
Hao Xu
Peking university
------------------ 原始邮件 ------------------ 发件人: "woshixuhao/PIC_code" @.>; 发送时间: 2022年10月9日(星期天) 下午4:49 @.>; @.***>; 主题: [woshixuhao/PIC_code] Ask for parametric PDE (Issue #1)
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>
— Reply to this email directly, view it on GitHub https://github.com/woshixuhao/PIC_code/issues/1#issuecomment-1272492677, or unsubscribe https://github.com/notifications/unsubscribe-auth/A2XEW3RAXSHJM7I4UYYNLN3WCKCFZANCNFSM6AAAAAARATKGEA . You are receiving this because you authored the thread.Message ID: @.***>
P.S. I am very happy to finally contact with you! I have tried to send several e-mails to your instituional mailbox, however, they are all rejected and sent back. I hope you can receieve this e-mail. I find that this reply mail cannot attach attachments, so you can add my QQ: 390260267 for further contact.
Hello, Thanks for your interest of our works! The code is attached in this e-mail. Here, I provide the case of parametric-convection-diffusion to help you better try this code. I am sorry that the code have not been sorted out well, so it may takes some time to get familiar with it.
The code is run as mentioned below:
run main.py to train a neural network, and the settings of network is in the neural_network.py
run GA-produce-meta-data.py to obtain the meta-data and then run 结合遗传PDE.py to conduct DLGA(structure). Here, you may manually adjust the domain of meta-data (line 16 and 17) to obtain the meta-data in different local windows and then run the 结合遗传PDE.py get the structure in this windows. After obtaining the structure in all windows, the optimal structure can be obtained as illustrated in the article.
run adaptive.py to get the coefficient of each term in each x or t. (This is a easier and faster way to obtain the coefficient, and the article use the PINN to optimize, but it is slow, if you want the PINN code, please tell me.)
run new-变系数.py to conduct DLGA(coefficient).
The middle result have been provided in the code, and you can debug anywhere. If you have any question ,please contact me at any time!
Hao Xu
Peking university
------------------ 原始邮件 ------------------ 发件人: "woshixuhao/PIC_code" @.>; 发送时间: 2022年10月9日(星期天) 下午4:49 @.>; @.***>; 主题: [woshixuhao/PIC_code] Ask for parametric PDE (Issue #1)
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>