.png){kind=icon}
Return
High performance algorithm framework for solving nonlinear PDEs by separable neural operators with wavelet methods
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ijnonlinmec.2025.105158.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• A framework of separable neural operator combining wavelet is proposed for efficiently and accurately solving nonlinear PDEs. • The wavelet hierarchies are leveraged to separate low/high frequency components for generating high-precision multi-fidelity datasets. • A hierarchical framework with independent neural operators for low-frequency patterns and high-frequency residuals is designed. • High-dimensional PDEs and integro-differential systems are solved to demonstrate the superior performance of the proposed method.
Journal
I
IF:
3.2
Papers:
398
Citations:
7.7K
Organization
No organization information available
