Abstract:Partial Differential Equations (PDEs) are an important branch of mathematics and are of great value in describing and solving natural phenomena and engineering problems. In recent years, with the rapid development of deep learning techniques, physics-informed neural networks (PINNs) have been successfully applied to solve partial differential equations and physical field simulations. Based on physical constraints, PINNs have received a lot of attention from researchers as they can quickly obtain highly accurate data-driven solutions with a small amount of data. In the realm of nonlinear science, the investigation of localized waves holds significant research significance and has found applications across various domains. In the present study, we have employed an enhanced version of physics-informed neural networks to conduct numerical simulations of localized waves governed by nonlinear partial differential equations. The improved PINNs not only incorporate the inherent constraints of the equations but also integrate constraints derived from gradient information. Moreover, we have employed an adaptive learning approach to dynamically update the weight coefficients of the loss function, thereby enhancing the training efficiency. In our experiments, we have chosen the nonlinear Schrödinger equation for the study and evaluated the accuracy of the localized wave simulation results by error analysis. The experimental results show that the improved PINNs significantly outperform the traditional PINNs with shorter training time and more accurate prediction results.

MetaPINNs: Predicting Soliton and Rogue Wave of Nonlinear PDEs via the Improved Physics-Informed Neural Networks Based on Meta-learned Optimization

Learning Rogue Waves of Nonlinear Schrödinger Equation with Enhanced Physics-Informed Neural Networks Simulator.

An Efficient Approximation Method Based on Enhanced Physics-Informed Neural Networks for Solving Localized Wave Solutions of PDEs

An Efficient Framework for Solving Forward and Inverse Problems of Nonlinear Partial Differential Equations Via Enhanced Physics-Informed Neural Network Based on Adaptive Learning

Boussinesq Equation Solved by the Physics-Informed Neural Networks

Solving Partial Differential Equations Using Deep Learning and Physical Constraints

RAR-PINN algorithm for the data-driven vector-soliton solutions and parameter discovery of coupled nonlinear equations

Explorations of certain nonlinear waves of the Boussinesq and Camassa–Holm equations using physics-informed neural networks

The improved backward compatible physics-informed neural networks for reducing error accumulation and applications in data-driven higher-order rogue waves

A Gradient-Enhanced Physics-Informed Neural Networks Method for the Wave Equation

Enhancing Accuracy of Physically Informed Neural Networks for Nonlinear Schrödinger Equations Through Multi-View Transfer Learning

Enhancing PINNs for solving PDEs via adaptive collocation point movement and adaptive loss weighting

Meta-PINN: Meta learning for improved neural network wavefield solutions

Physics Informed Neural Networks (PINNs) as intelligent computing technique for solving partial differential equations: Limitation and Future prospects

Pre-training strategy for solving evolution equations based on physics-informed neural networks

Hypernetwork-based Meta-Learning for Low-Rank Physics-Informed Neural Networks

Physics-Informed Neural Networks (PINNs) for Parameterized PDEs: A Metalearning Approach

Numerical Solver-Independent Seismic Wave Simulation Using Task-Decomposed Physics-Informed Neural Networks.

A metalearning approach for Physics-Informed Neural Networks (PINNs): Application to parameterized PDEs

A two-stage physics-informed neural network method based on conserved quantities and applications in localized wave solutions

Physics-Informed neural network solver for numerical analysis in geoengineering