Dongfu Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Dongfu Liu has a total of 3 co-authors. Here are the top co-authors by number of shared publications:
- Fei Luo
College of Mathematics and Statistics, Sichuan University of Science and Engineering, Zigong 643000, China
1 shared publication - Yiran Zhang
School of Mathematics and Statistics, Sichuan University of Science and Engineering, Zigong 643000, P. R. China
1 shared publication - Yundong Li
School of Mathematics and Statistics, Sichuan University of Science and Engineering, Zigong 643000, P. R. China
1 shared publication
Latest Publications
- 28 October 2026
- Published by World Scientific Pub Co Pte Ltd in International Journal of Computational Methods
Warm-Up-Calibrated Adaptive Weighting for Spatiotemporal Residual-Derivative Balance in Gradient-Enhanced PINNs
Neural-network methods are widely used for partial differential equations (PDEs). Within this framework, gradient-enhanced physics-informed neural networks (gPINNs) incorporate first-order residual derivatives, while fixed directional weighting may unbalance spatial and temporal residual-gradient contributions. Accordingly, an adaptive gradient-balanced gPINN (AGB-gPINN) is proposed using warm-up calibration, dimensionless relative states, and a bounded fixed-sum directional budget without trainable weighting parameters. For the Burgers equation, median relative [Formula: see text] error and PDE-MSE over nine seeds decrease by 41.9% and 57.0% relative to fixed-weight gPINN; the Allen–Cahn equation shows smaller gains and lower variability. These results support more balanced and stable gPINN training.
- 28 October 2026
- Published by World Scientific Pub Co Pte Ltd in International Journal of Computational Methods
Warm-Up-Calibrated Adaptive Weighting for Spatiotemporal Residual-Derivative Balance in Gradient-Enhanced PINNs
Neural-network methods are widely used for partial differential equations (PDEs). Within this framework, gradient-enhanced physics-informed neural networks (gPINNs) incorporate first-order residual derivatives, while fixed directional weighting may unbalance spatial and temporal residual-gradient contributions. Accordingly, an adaptive gradient-balanced gPINN (AGB-gPINN) is proposed using warm-up calibration, dimensionless relative states, and a bounded fixed-sum directional budget without trainable weighting parameters. For the Burgers equation, median relative [Formula: see text] error and PDE-MSE over nine seeds decrease by 41.9% and 57.0% relative to fixed-weight gPINN; the Allen–Cahn equation shows smaller gains and lower variability. These results support more balanced and stable gPINN training.